Upload
dinhhanh
View
219
Download
0
Embed Size (px)
Citation preview
TESIS DOCTORAL
MECANISMOS DE CONTROL COGNITIVO Y EMOCIONAL:
PROCESOS CONSCIENTES Y NO CONSCIENTES
Doctoranda: Mª Asunción Panadero Sanchis
Director: Pío Tudela Garmendia
Programa Oficial de Posgrado en Diseños y Aplicaciones en
Psicología y Salud (P33.56.1)
NOVIEMBRE 2015
Editor: Universidad de Granada. Tesis DoctoralesAutora: María Asunción Panadero SanchísISBN: 978-84-9125-555-0URI: http://hdl.handle.net/10481/42777
El doctorando Mª Asunción Panadero Sanchis y el director de la tesis Pío Tudela Garmendia, garantizamos, al firmar esta tesis doctoral, que el trabajo ha sido realizado por el doctorando bajo la dirección de los directores de la tesis y hasta donde nuestro conocimiento alcanza, en la realización del trabajo, se han respetado los derechos de otros autores a ser citados, cuando se han utilizado sus resultados o publicaciones.
Firmado en Granada a día 5 de Noviembre de 2015
Pío Tudela Garmendia Mª Asunción Panadero Sanchis
:
AGRADECIMIENTOS
…Cuento en mi vida con personas de un valor incalculable….
Como no agradecer a todas aquellas personas que me han acompañado a lo largo de todo este largo y duro proceso. No puedo olvidar a aquellos y aquellas que me han brindado su apoyo, que me han arrastrado a la calle, a los que yo he arrastrado a la calle, a los
que me han hecho reír, cuando lo que me apetecía era llorar o gritar. En fin a todos los que, cuando lean esto, ya sabrán quienes son.
Tampoco puedo olvidar a todos aquellos compañeros y compañeras que me han regalado su tiempo y sus conocimientos y que me han ayudado en lo que he necesitado para que esta tesis sea lo que es
hoy.
Sin embargo, aunque han sido muchos y muchas, y no me gustaría olvidar a nadie, debo destacar a algunas de esas personas, que son tan parte de esta tesis como yo misma, o incluso más, porque sin
ellas esto no habría sido posible.
Pio, bueno, la verdad es que tengo tanto que agradecerte que no se ni por dónde empezar. Hace ya unos años que nos
embarcamos juntos en esto y nunca has dejado de creer en mi ni de darme tu apoyo. Gracias por tu infinita paciencia, por tu ayuda, por tu escucha, en momentos difíciles. Para mí ha significado mucho. Gracias por tus infinitos conocimientos, que han dado esta tesis la
forma que tiene. Sin ti no habría sido posible. Gracias por tu tiempo y por tu trabajo, de un valor incalculable y por dedicarle a esto tanto
o más que yo. Para mí siempre serás más que mi director de tesis, ¡Muchas gracias!
Gracias a mi familia que son las personas que han hecho que hoy por hoy sea como soy y que no me rinda bajo ninguna
circunstancia y que termine siempre lo que he empezado. Gracias mamá por soportar mis malos humores y mis locuras transitorias.
Has sabido entender que eran producto de lo externo y las has capeado como solo tú sabes hacerlo. ¡Te quiero!
Como no poner en un lugar privilegiado a quien, aunque no de sangre, ha sido mi hermana durante todos estos años. Como digo,
mi hermana, mi amiga, mi compañera de piso durante más de 6 años. Ella ha vivido de cerca esta etapa de mi vida. Cuantos días de aguantarme despotricar, de oír que no lo iba a conseguir, de verme
llegar tarde del laboratorio de pasar participantes, y de verme desesperar cuando tenía que tirar muchos de ellos por lo malditos
parpadeos. Cuantos momentos de risas (sobre todo los desayunos), momentos que hacían que todo fuera más llevadero y que no me dejara arrastrar por la desesperación en determinadas situaciones.
Sabes que te quiero ¡Gracias!
Conchi querida, ¿qué te digo que no te haya dicho ya? Esta tesis es tan tuya como mía, lo sabes. No hay palabras ni combinaciones de letras que expresen bien lo que te agradezco todo lo que has hecho por mí durante todo este tiempo. Créeme cuando te digo que sin ti
esto no habría salido nunca a flote. Además de por todo esto, gracias por los cafés en Psicología primero y luego en Farmacia,
oliendo a tostada quemadilla, con ese sueño profundo que solo da el laboratorio, con esa luz tenue. Gracias por hacerme remontar cada
vez que flaqueaba (han sido tantas…). ¡Gracias por todo!
Kitty, en fin, yo sé que no necesito escribirte nada porque me lees la mente, pero aun así déjame decirte que esos momentos por todos los rincones de Granada destrozándonos el hígado no han tenido precio.
Y tampoco nuestras charlas hasta las mil de la mañana sobre toda clase de temas y en muchas ocasiones sobre esas tontería que tú y
yo compartimos y que nadie más entiende. ¡Te quiero!
Marifé, amore tú sabes lo que yo te quiero y lo que te agradezco todos los momentos que hemos pasado. No sabes hasta qué punto
me han salvado en momentos difíciles.
Quique, cómo no agradecerte los momentos frikis que me han hecho desconectar siempre que lo he necesitado, por nuestros cafés en ese sitio que nunca me acuerdo de cómo se llama y al que nunca se llegar por mí misma. Gracias por las noches de cervezas y tapas y
algún que otro gin tonic. ¡Gracias por soportarme cuando ni yo misma me soportaba!
Toñi, gracias por estar ahí cuando te he necesitado. Por sacarme y por rescatarme y por darme alguna que otra colleja
mental cuando me ponían inaguantable y, por no rendirte y esperar hasta que estuve preparada para decir basta.
Eduardo, gracias también por todas nuestras charlas y discusiones sobre la verdad de la ciencia en psicología. Algún día llegaremos a
alguna conclusión lógica. Gracias por “reírte” de mis neuras y quitarles importancia y por darme siempre buenos consejos.
Gracias a todos aquellos compañeros y compañeras que han dedicado tiempo a escuchar mis problemas y mis dudas en los
momentos que más los he necesitado. El que alguien te dedique tiempo siempre es algo que valoro por encima de muchas cosas.
Gracias a mi familia del teatro. Vosotros hacéis que mi mente se evada de todo hasta el punto del reseteo. Por supuesto, como no,
gracias a mis niñas y a nuestros viernes que ya se han convertido en tradición.
Gracias a mi gente de los caballos, por hace que disfrute además de mi afición, de buenos momentos entre muy buena gente. Os quiero
¡Gracias a todos!
INDICES
ÍNDICE DE CONTENIDO 11 ÍNDICE DE FIGURAS 13 INTRODUCCIÓN 17 1. Contextualización histórica 19
1.1. Antecedentes histórico 19 1.2. Dos modelos clásicos acerca de cómo la atención modula
el procesamiento controlado y el procesamiento automático 23 1.2.1.Norman y Shallice (1980) 23 1.2.2. Posner y Petersen (1990) 27
2. Contexto reciente en el que se enmarca nuestra investigación 35 2.1. La Hipótesis de Monitorización del Conflicto 37 2.2. Evidencias a favor de la Teoría de Monitorización del
Conflicto 38 3. Efectos de adaptación al conflicto 45
3.1. Efectos secuenciales: Efecto Gratton 46 3.2. Efectos de la frecuencia del tipo de ensayo 48
3.2.1. Efecto de proporción de congruencia (PCE) 49 3.2.2. Efecto de proporción de congruencia específico
del ítem (ISPCE) 51 3.2.3. Context-Specific Compatibility Effect 54
4. Aclarando el concepto de consciencia 57 5. Nuestra investigación 62 SERIE EXPERIMENTAL I-EXPERIMENTO 1 65 1. Introduction 68 2. Material and methods 78
2.1. Participants 78 2.2. Stimuli and apparatus 78 2.3. Procedure and design 79
2.3.1. Procedure 79 2.3.2. Design 81
2.4. EEG recording and data analysis 82 3. Results 85
3.1. Behavioral results 85 3.1.1. Threshold setting 85 3.1.2. RT analysis 85
3.2. Electrophysiological data 86 3.2.1. Target analysis 86 3.2.2. Mask analysis 90
4. Discussion 91
ÍNDICES
12
SERIE EXPERIMENTAL II-EXPERIMENTO 1 101 1. Introduction 103 2. Material and methods 111
2.1. Participants 111 2.2 Stimuli and apparatus 112 2.3. Procedure and design 112
2.3.1. Procedure 112 2.3.2. Design 114
2.4. EEG recording and analysis 114 3. Results 117
3.1. Behavioral analysis 117 3.1.1. RT analysis 117
3.2. Electrophysiological analysis 119 3.2.1. Bilateral group of electrodes 119 3.2.2. Centro-parietal group of electrodes 122
4. Discussion 127 SERIE EXPERIMENTAL II-EXPERIMENTO 2 137 1. Introduction 139 2. Material and methods 151
2.1. Participants 151 2.2 Stimuli and apparatus 152 2.3. Procedure and design 153
2.3.1. Procedure 153 2.3.2. Design 154
2.4. EEG recording and analysis 155 3. Results 158
3.1. Behavioral analysis 158 3.1.1. Threshold setting 158 3.1.2. RT analysis 158
3.2. EEG analysis 161 4. Discussion 176 DISCUSIÓN GENERAL 187 El efecto de adaptación al conflicto 192 El procesamiento no consciente 200 Conclusiones 205 REFERENCIAS 209
FIGURAS
13
FIGURAS
Figura 1.1. Modelo de Control Ejecutivo de Norman y Shallice ........... 24 Figure 2.1. (A) Sequence of events in an incongruent trial. Words have been translated to English. (B) Groups of electrodes selected for P2 (circles), P3 (squares) and N2 (rhombuses) components. Topographies of the average temporal windows of analysis show the maximum level of amplitude for each group of electrodes. ... 81 Figure 2.2. (A) Interaction of Incongruency percentage x Congruency effect on RT. (B) Waves and topographies of the main effect of Incongruency percentage on the P2 amplitude. (C) Waves and topographies corresponding to the interaction of Incongruency percentage x Congruency on the P3 amplitude. The 33% effect on Congruency appears on the left (33% congruent in gray and 33% incongruent in black) while the 66% effect on Congruency appears on the right (66% congruent in gray and 66% incongruent in black). In the center we present the amplitude table for this interaction. ... 87 Figure 2.3. Waves and topographies corresponding to the interaction of Incongruency percentage x Congruency on the N2 amplitude. The 33% effect on Congruency appears on the left (33% congruent in gray and 33% incongruent in black) while the 66% effect on Congruency appears on the right (66% congruent in gray and 66% incongruent in black). In the center we present the amplitude table for this interaction. .................................................... 89 Figure 3.1. Sequence of a non-emotional congruent and emotional incongruent trial. .................................................................................... 113 Figure 3.2. Group of electrodes selected for the analysis and topographies showing amplitude differences in non-emotional and emotional tasks. . 116 Figure 3.3. Behavioral results: interaction effect Task x Current Trial Congruency ............................................................................................ 118 Figure 3.4. Average waves of the congruent and incongruent trials when previous trial was congruent or incongruent for the left hemisphere (in the left) and right hemisphere (in the right). ................................................ 120 Figure 3.5. Averages representation of the interaction effect Previous Trial Congruency x Current Trial Congruency ...................................... 121
ÍNDICES
14
Figure 3.6. Average waves of the congruent and incongruent trials for the non-emotional (A) and emotional (B) tasks and for the previous congruent (left) and previous incongruent (right) trials. ......................................... 122 Figure 3.7. Graphic representation of the interaction effect between Task x Current Trial Congruency. .................................................................. 124 Figure 3.8. N2 Average amplitude of the congruent and incongruent trials for the non-emotional (A) and emotional (B) tasks and for the previous congruent (left) and incongruen trials (right). ........................................ 125 Figure 3.9. P3 average amplitude of the congruent and incongruent trials for the non-emotional (A) and emotional (B) tasks and for the previous congruent (left) and previous incongruent (right) trials. ........................ 126 Figure 4.1. (A) Trial sequence of events. (B) Groups of electrodes selected for N170 (triangles), P2 (circles), P3 (squares) and N2 (rhombuses) components........................................................................ 157 Figure. 4.2. Interaction effect Incongruency percentage x Current trial congruency in the RT. ............................................................................ 159 Figure 4.3. Interaction effect Incongruency percentage x Previous trial congruency x Current trial congruency in the RT. ................................. 160 Figure 4.4. Waves and graphic representation of the latency interaction effect Incongruency percentage x Current trial congruency effect in N170 latency. ................................................................................................... 161 Figure 4.5. Waves and graphic representation of the interaction effect Previous Congruency x Congruency in N170 latency. .......................... 162 Figure 4.6. Graphic representation of the interaction effect Incongruency percentage x Previous trial congruency x Current trial congruency for the N170 latency. ......................................................................................... 163 Figure 4.7. Graphic representation of the interaction effect Task x Incongruency percentage x Previous trial congruency in the N170 latency. ................................................................................................................ 164 Figure 4.8. Topographies corresponding to P2 peak (240 ms) .............. 165 Figure 4.9. Waves and graphic representations of the interaction effect Task x Previous trial congruency x Current trial congruency for the P2 amplitude. ............................................................................................... 167 Figure 4.10.: Topographies corresponding to the N2 component. ........ 169 Figure 4.11. Waves and graphic representations of the Interaction effect Previous trial congruency x Current trial congruency for the N2 amplitude. ............................................................................................... 170 Figure 4.12. Topographies corresponding to P3 peak (380 ms) ............ 171
FIGURAS
15
Figure 4.13. Waves and Graphic representations of the close to significance interaction effect Previous trial congruency x Current trial congruency for the P3 amplitude. .......................................................... 172 Figure 4.14. Waves of the second order interaction effect Incongruency percentage x Previous trial congruency x Current trial congruency in emotional task for P3 amplitude. ............................................................ 174 Figure 4.15. Graphic representation of the second order interaction effect Incongruency percentage x Previous trial congruency x Current trial congruency in emotional task for P3 amplitude. .................................... 175 Figure 4.16. Interaction effect Previous trial congruency x Current trial congruency in P3 amplitudes for the Non-Emotional task for P3 amplitude. ............................................................................................... 175
Capítulo 1
INTRODUCCIÓN
1.- CONTEXTUALIZACION HISTORICA
1.1. Antecedentes previos
A partir del siglo XIX es cuando empezamos a encontrarnos
grandes conceptualizaciones de la atención como: mecanismo
selectivo, asignación de capacidad y lo que va a llevar a toda una
línea de estudio dentro de la cual se va a encuadrar nuestra
investigación, que es la distinción entre un tipo de atención más
voluntaria y controlada, a la cual se le atribuye la necesidad de
consciencia y otro tipo de atención, en la que el individuo tiene
menos control, denominada automática y que por ende se ha
asumido que no necesita de la consciencia, aunque tampoco ésta la
perjudica. La idea recurrente es la necesidad de distinguir aquellos
estímulos sobre los que no se atiende y aquellos sobre los que sí, y
que por tanto son conscientes. Así Wilhelm Wundt distingue entre
percepción y apercepción, siendo esta última la que está constituida
por aquellos estímulos que se encuentran en el foco de la
consciencia, debido a un proceso de focalización atencional, de
manera que la atención se empieza a concebir como una actividad
interna que determina los grados de consciencia de la información
percibida.
El control cognitivo puede definirse como una función que
nos permite alcanzar con éxito nuestras metas, mediante la
adaptación flexible de nuestro comportamiento al medio,
potenciando acciones que se dirigen hacia nuestras metas y/o
conteniéndolas para evitar resultados no deseados. Para que esto se
CAPÍTULO 1
20
lleve a cabo es necesaria la coordinación de múltiples sistemas. Así
la memoria es necesaria para mantener nuestra meta, la atención
para seleccionar la información relevante de entre toda la
información irrelevante del medio y la percepción, para procesar
dicha información. El control cognitivo atencional es crucial para
nuestro comportamiento, ya que sin él, ninguna de nuestras
actividades de la vida diaria podrían llevarse a cabo.
Para poder llegar a comprender bien el punto en el que esta
tesis está enmarcada hay que hacer un repaso a cerca de cómo se ha
llegado a la definición de control tal y como hoy en día se conoce.
Dado que el tema principal de este trabajo es el interés en mostrar la
existencia de procesos de control no conscientes debemos
detenernos y echar un ojo a las distinciones que, a lo largo de la
historia de la Psicología cognitiva, se han hecho entre
procesamiento controlado-consciente y procesamiento automático-
no consciente.
Si observamos cualquier actividad de las que realizamos
todos los días (lavarnos los dientes, vestirnos, ir hacia el trabajo,
levantarnos de la cama, etc) y “prestamos atención” a cada una de
ellas, nos daremos cuenta de que la mayoría son para nosotros
actividades que podríamos considerar automáticas. Por efecto de la
práctica hemos llegado a dominar tanto cada uno de los pasos que
realizamos para llegar a la consecución de cada una de estas tareas,
que simplemente ya no somos conscientes de ellos a menos que les
pongamos atención de manera voluntaria. Sin embargo, cuando
tenemos que enfrentarnos por ejemplo a una situación novedosa
para nosotros, enseguida nos damos cuenta de la necesidad de poner
INTRODUCCIÓN
21
el máximo de nuestros recursos atencionales sobre dicha tarea para
poder llevarla a cabo lo mejor posible. Es por esta razón por la que
se ha asociado consistentemente el control voluntario con la
consciencia.
En la psicología moderna, la primera aportación acerca de la
distinción entre estos dos tipos de procesamiento fue realizada por
Posner y Snyder en 1975. Hasta este momento la línea que
diferenciaba ambos procesos era bastante difusa, de manera que lo
que estos autores pretendieron fue dilucidar con más claridad las
diferencias entre ambos tipos de procesamiento. Para ellos un
proceso automático era aquel que ocurría sin atención, sin
consciencia y sin provocar ninguna interferencia sobre cualquier
otra tarea que se estuviera realizando de manera simultánea. En
estos casos el procesamiento que se realizaría sobre la información
sería en paralelo, a diferencia de aquellas situaciones en las que se
requieren procesos controlados y conscientes y que estarían guiadas
por un procesamiento serial, de manera que para que no se
produjera una interferencias las tareas tendrían que llevarse a cabo
una detrás de otra.
Se ha considerado que la práctica sistemática de una tarea
lleva al aprendizaje de ésta, y es mediante ese aprendizaje que se
consigue la automatización de las mismas, por lo que pueden ser
realizadas a la vez que podemos dirigir nuestra atención a otra cosa
(Shiffrin & Schneider, 1977; Schneider & Shiffrin, 1977). Sin
embargo, existe una contrapartida: una vez que una acción se ha
automatizado hasta este punto, se pierde flexibilidad en cuando a
ejecución se refiere, por lo que cuando queremos introducir un
CAPÍTULO 1
22
cambio en la ejecución de dicha tarea y modificar esa acción
practicada masivamente, debemos realizar un gran esfuerzo que
lleva a la consecución de numerosos errores (Schneider & Shiffrin,
1977). Es esto lo que ocurre con el conocido como efecto Stroop.
En este caso compite una tarea muy bien aprendida como es la
lectura de una palabra, con otra que es novedosa y que requiere de
más esfuerzo como es el nombramiento del color de la tinta en la
que está escrita dicha palabra. Así cuando la palabra nombra el
mismo color en el que está impresa la competencia entre la tarea
automática lectura y la tarea controlada nombrar el color será
menor que si la palabra nombra otro color diferente al que está
impresa (Stroop, 1935). En el primer caso hablaríamos de una
condición congruente, mientras que en el segundo caso hablaríamos
de una condición incongruente.
De la misma manera que con las tareas automáticas, se han
tratado de tipificar las situaciones en las que se requiere de control
consciente. Norman y Shallice (1980), proponen cinco situaciones
claves en las que es necesario del control consciente con el fin de
maximizar la eficacia de la ejecución. Estas serían: situaciones
novedosas, situaciones que entrañan peligro o riesgo para el
individuo, situaciones difíciles, situaciones en las que se ha
cometido un error y por último, la ya mencionada situación en la
que es preciso inhibir una respuesta automática que es inapropiada
o inútil para la consecución de las metas actuales de la persona.
INTRODUCCIÓN
23
1.2. Dos modelos clásicos que pretenden explicar cómo la atención
modula el procesamiento controlado y el procesamiento
automático: Modelo de Norman y Shallice (1980) y Teoría de las
Redes Atencionales de Posner y Snyder (1990).
Surge cada vez más la necesidad de explicar el papel que la
atención tiene en el control de la acción, ya sea una acción externa o
una acción interna en la que solo intervenga el procesamiento
cognitivo. Y no solo la necesidad de explicar el control de la acción
en general, sino de diferenciar y abordar el cómo la atención
modula las acciones automáticas y aquellas que están sujetas a un
control deliberado y consciente.
1.2.1 Modelo de Norman y Shallice
Su principal objetivo a la hora de formular un modelo de la
atención es explicar cómo ésta va a influir en el control de las
acciones, tanto las que son realizadas de manera automática como
las que se llevan a cabo bajo control voluntario. El marco teórico
que ellos proponen se basa en una serie de esquemas activos que se
organizan como parte de una secuencia de acción determinada.
Estos esquemas permanecerán a la espera del set de acción
apropiado, de manera que una vez que dicho set se produzca estos
esquemas puedan ser seleccionados con el fin de controlar la
acción. Norman y Shallice van a tomar el término automático en el
sentido de aquellas acciones que no van a generar interferencia con
otras tareas. Mientras que para ellos existen en general cinco
CAPÍTULO 1
24
categorías de acciones que si necesitarían control voluntario y por lo
tanto la consciencia, y que serían las que hemos mencionado con
anterioridad. En estas situaciones la aplicación no controlada de un
esquema puede llevar a la comisión de errores o a una situación que
no es deseable.
Figura 1.1. Modelo de Control Ejecutivo de Norman y Shallice
El aspecto central del modelo se basa en que todo el
comportamiento humano está mediatizado por esquemas mentales,
que son los que van a interpretar las señales que llegan al sistema de
procesamiento de la información y los que van a seleccionar la
respuesta adecuada en cada momento. Ante la pregunta de cómo
esos esquemas interaccionan entre sí y como se llega a la elección
de la respuesta apropiada a cada situación, proponen dos
mecanismos fundamentales: el contention scheduling (CS) y el
Sistema Atencional Supervisor (SAS).
INTRODUCCIÓN
25
a) Contention Scheduling:
Como hemos comentado anteriormente el eje central del
modelo de Norman y Shallice es la existencia de esquemas de
acción que deberán ser activados para que dicha acción se lleve a
cabo. Para ellos el mecanismo de CS sería aquel que evalúa la
relevancia relativa de cada uno de los esquemas y por tanto, va a
gestionar el comportamiento en función de dicha evaluación
evitando así la competición en la selección de esquemas o
favoreciendo la cooperación entre los mismos para potenciar una
determinada acción. Podríamos resumir el funcionamiento de este
sistema como la evaluación del valor de una serie de sets de
esquemas que están compitiendo entre sí, con el fin de decidir cuál
de ellos es el más adecuado en función de la acción que se quiere
llevar a cabo.
Este sistema funciona especialmente bien cuando tratamos
con acciones rutinarias complejas. De esta forma un estímulo
ambiental va a desencadenar una conducta. Para que dicha conducta
sea la apropiada en cada caso será necesario un sistema de
inhibición recíproca, de manera que el esquema de acción más
activo será aquel que prevalezca, mientras los demás se mantienen
inactivos temporalmente, Como ya hemos dicho, este sistema será
especialmente útil en el caso de que nos enfrentemos con acciones
muy rutinarias y bien especificadas. En ausencia de alguna
estimulación ambiental este sistema quedará inactivo o latente.
CAPÍTULO 1
26
Pero, ¿qué ocurre cuando se presenta un estímulo o una
situación para la que no hay una acción rutinaria especificada o ésta
no es apropiada? Es en estos casos, cuando se necesita de un
sistema que module y controle al CS y que favorezca la toma de
decisiones o la inhibición de la respuesta rutinaria inapropiada. Es
ahí donde entra en escena el Sistema Atencional Supervisor (SAS).
b) El SAS:
El SAS va a estar activo y en pleno funcionamiento en
cualquiera de las cinco situaciones que se han mencionado
anteriormente como generadoras de la necesidad de control
cognitivo y atencional. Este sistema puede modificar las fuerzas de
acción rivales dando más preponderancia a unas o a otras en
función de las demandas de la nueva situación.
Asímismo, ante aquellas situaciones que no generan un
modelo de acción determinado, el SAS puede generar uno que se
adecue a dicha situación. De esta manera podemos decir que el SAS
puede inhibir una respuesta automática y perseverante, así como
puede generar conductas nuevas ante aquellas situaciones que no
elicitan secuencias de acción rutinarias.
La acción de control del SAS sería, por tanto, indirecta ya
que éste solo controla y regula los niveles de activación o inhibición
de los esquemas y no su selección en sí. Es decir, este sistema
aumentaría o disminuiría deliberadamente el valor que un esquema
determinado posee, según la acción que se realice o las demandas
de la tarea. Pero sería el CS el encargado de seleccionarlo para
INTRODUCCIÓN
27
aplicarlo a la acción. A la influencia que el SAS ejerce sobre los
niveles de activación de los esquemas de acción de CS hay que
añadirle además los factores motivacionales de cada individuo. La
motivación sería un factor que iría actuando de manera lenta y
progresiva a lo largo del tiempo sobre el CS para dirigir dichas
estructuras a la consecución de una meta más a largo plazo,
mediante la activación de los esquemas relevantes y la selección de
los correspondientes.
1.2.2 La Teoría Atencional de Posner y Petersen (1990)
Al igual que Norman y Shallice, Posner intenta dar cuenta
de todos los aspectos fundamentales de la atención mediante ka
propuesta de una teoría integradora (Posner y Boies, 1971; Posner y
Petersen, 1990; Posner y Rothbart, 1991; Posner y Dehaene, 1994).
En un primer momento, para Posner y Boies (1971) la atención
tenía tres componentes fundamentales, necesarios para entender
como ésta influía en las acciones: el primero de esos componentes
era la alerta, entendida como un estado de vigilancia o de atención
sostenida que se mantenía durante un periodo largo de tiempo para
poder realizar una determinada tarea o actividad. El segundo
componente hacía referencia a la selección de información. Por
último el tercer componente hace referencia a la capacidad de
procesamiento.
Posteriormente, a medida que estudiaron la relación entre
estas funciones con las estructuras cerebrales que las sustentan, la
teoría fue evolucionando hasta llegar a la formulación de aquella
CAPÍTULO 1
28
que postula la existencia de varias redes atencionales separadas pero
relacionadas entre sí, que serían: la Red Atencional de Vigilancia o
Alerta, la Red Atencional Posterior o de orientación y la Red
Atencional Anterior o del Control Ejecutivo.
a) Red Atencional de Vigilancia o Alerta
Esta es la red encargada de que el sistema atencional se
mantenga con un nivel de activación y de ejecución óptimo durante
la realización de una tarea que es el que va a permitir que se puedan
llevar a cabo las distintas actividades y tareas de la vida con
eficacia. Esta red no solo va a encargarse de este tipo de alerta que
podría considerarse tónica y mantenida durante periodos largos de
tiempo, sino que también tiene la función de mantener un tipo de
alerta más fásica y de corta duración ante la presencia de una señal
de aviso que indique la presencia inminente, de un estímulo
esperado. Todo esto es muy útil a la hora de acelerar la respuesta
del organismo ante determinadas situaciones, pero también tiene un
inconveniente y es que al ganar en rapidez se pierde un poco en
precisión y se aumenta la comisión de errores y/o anticipaciones,
por tanto solo nos prepara para la acción pero no mejora el nivel del
procesamiento de los estímulos (Posner, Klein, Summers y Buggie,
1973; Posner, 1978).
Tradicionalmente se han utilizado metodologías diferentes
para estudiar ambos tipos de alerta. En el caso del estudio de la
alerta fásica, el paradigma tradicional consiste en la presentación
de una señal de aviso justo antes de la aparición de un estímulo
INTRODUCCIÓN
29
objetivo al cual hay que responder, que como ya hemos comentado
anteriormente produce un estado de preparación para la detección
del dicho estímulo, y mejora el TR en la respuesta a éste, no porque
aporte más información acerca de sus características, sino porque
facilita la respuesta de orientación hacia el mismo y por tanto
facilita su percepción.
Con respecto a la alerta tónica, los estudios que se han
realizado han estado marcados por las fluctuaciones que en ésta se
producen a lo largo del día (ritmos circadianos), y por las variables
que influyen en dichos cambios (temperatura corporal, secreción de
cortisol, etc). Las tareas utilizadas en dichos estudios comparten las
características de ser largas y aburridas con el fin de poder medir las
fluctuaciones en atención sostenida. Se ha visto que el rendimiento
en estas tareas y el TR son mejores durante las horas más tempranas
de la mañana, mientras que dicho rendimiento va cayendo a lo largo
del día para volver a aumentar durante la noche (Posner, 1975).
Estudios de lesión y neuroimagen han situado esta red
lateralizada en el hemisferio derecho en estructuras que incluirían el
córtex frontal y parietal derecho, donde se reciben innumerables
proyecciones noradrenérgicas procedentes de Locus Coeruleus
(Aston-Jones y Cohen, 2005). Otros estudios de neuroimágen (Sturn
y Willmes, 2001) han ampliado esta información mostrando que
numerosas áreas talámicas también intervienen, tanto en estados de
alerta fásica, como en estados de alerta tónica o vigilancia.
CAPÍTULO 1
30
b) Red Atencional Posterior o de Orientación
Sería aquella encargada de dirigir la atención hacia un
determinado estímulo que bien posee características peculiares que
lo hacen único, es nuevo o bien aparece de manera abrupta (Ruz y
Lupiáñez, 2002). De esta manera se prioriza un imput sensorial bien
por su modalidad o por su localización (Posner y Petersen 1990). Se
ha observado un efecto de facilitación en la percepción cuando la
atención está orientada hacia un determinado estímulo, aunque esta
señal no prediga el lugar exacto en el que dicho estímulo va a
aparecer (Posner, 1980; Posner y Cohen, 1984). Para el estudio del
efecto de orientación se han utilizado tareas de señalización
(Posner, Inhoff, Friedrich y Cohen, 1987). Durante este tipo de
tareas, para producir el efecto de orientación, se utilizan dos tipos
de señales, que pueden bien indicar dónde va a aparecer el estímulo
objetivo (Marrocco y Davidson, 1998), o bien aparecer en el mismo
sitio que el estímulo objetivo: el primer tipo de señales, serían las
llamadas centrales. Este tipo de señales, en líneas generales son
predictivas, es decir, van a dar información del lugar en el que va a
aparecer el estímulo objetivo, por lo que la persona va a poner
orientar su atención hacia ese lugar. El segundo tipo de señales,
serían las periféricas. En este caso, aunque la señal aparece en un
posible lugar de aparición del estímulo objetivo, no son predictivas
(el estímulo aparecerá en la posición que indica la señal el 50% de
las veces).
Numerosos estudios en humanos y monos han mostrado que
los moduladores de la Norepinefrina afecta a la alerta pero no a la
INTRODUCCIÓN
31
orientación, si bien al contrario, aquellas drogas moduladoras del
neurotransmisor acetilcolina, afecta al efecto de orientación pero no
al de alerta, sugiriéndose así la existencia de una independencia
clara entre ambos procesos (Fernández-Duque y Posner, 1977).
Pese a esta independencia, ambos procesos pueden funcionar a la
vez y potenciar sus efectos, ya que en la vida diaria la mayoría de
las situaciones de aviso arrojan información acera del cuándo y del
dónde va a ocurrir algo (Fan et al., 2009).
Neuroanatómicamente situamos esta red en estructuras tales
como el córtex parietal posterior, los núcleos pulvinar y reticular del
Tálamo y los colículos superiores. Actualmente, los estudios de
neuroimagen arrojan datos a favor de la intervención en procesos de
orientación de áreas frontales (Frontal Eye Field) y de áreas
posteriores (Corbetta, et al., 1998; Thompson et al., 2005), situadas
en el surco intraparietal. En lo que respecta a las áreas parietales, se
ha visto su implicación en diversas formas de procesamiento.
Determinadas áreas están más especializadas en procesamiento
motor, traducido en el movimiento directo de los ojos hacia la señal,
mientras que otras zonas se especializarían más en un tipo de
procesamiento más sensorial, en el que la orientación de la atención
no sería tan claramente visible, pero sin embargo estaría igualmente
presente. Este sistema constituye un mecanismo más dorsal que se
activa por ejemplo cuando se presenta una flecha que hace de señal
y que va a predecir el lugar de aparición del target (señal central).
La función que se le ha atribuido a este sistema es la de un rápido
control estratégico sobre la atención.
CAPÍTULO 1
32
Cuando la señal es periférica y no predice el lugar de
aparición del objetivo, puede ocurrir que la atención se oriente hacia
un lugar y que el estímulo no aparezca ahí, por lo que deba haber un
cambio en la orientación de la atención. Ese cambio se asocia con
activación del la unión temporo-parietal y del cortex frontal ventral
y se asocia con la interrupción de la señal para que se pueda
producir ese cambio atencional. Esto formaría parte de un sistema
más ventral, que se activaría, no tras la señal como el dorsal, sino
tras el estímulo objetivo y se identificaría como una parte de la red
encargada de procesar los sucesos sensoriales.
c) Red Atencional Anterior o de Control Ejecutivo
Esta red se ha asociado a la gestión del control voluntario de
la acción en aquellas situaciones novedosas, en resolución de
conflicto tanto entre estímulos como entre respuestas, en situaciones
que requieren de algún tipo de planificación o estrategia, etc
(Posner y Raichle, 1994). Otra de las funciones que se le ha
atribuido a esta red es la de detección consciente de los estímulos
(Posner y Raichle, 1994) a la vez que procesos de memoria de
trabajo (Posner y Dehaene, 1994) y en situaciones de recompensa
(Hampton y O´Doherty, 2007).
Lejos de considerar a esta red como una subparte de otros
sistemas que regulan estos procesos, Posner y Petersen (1990)
consideraron a esta red como un sistema que englobaba procesos de
focalización atencional, así como de regulación de las demás redes
de procesamiento. En su revisión más actual (Petersen y Posner,
INTRODUCCIÓN
33
2012) este sistema tendría como función principal la regulación top-
down, por lo que estaría estrechamente relacionada con el control
ejecutivo, el cual ha sido asociado al procesamiento consciente de la
información.
Estos autores ponen el énfasis en una estructura cerebral que
para ellos es clave en el funcionamiento de esta red: el cortex
cingulado anterior (ACC). Sin embargo, el papel del ACC en lo que
se refiere al control ejecutivo es una cuestión que aún se debate
entre dos grandes teorías. Una de ellas considera esta estructura
clave en la monitorización del conflicto, en constante comunicación
con áreas frontales laterales que serían las encargadas de resolver
dicho conflicto (Botvinick et al., 2001; Carter y Krug, 2012). La
otra gran teoría es la que va a favor de la existencia de dos redes
diferentes de control top-down (Dosenbach et al., 2008). A estos
sistemas llegarían tres tipos de señales con el fin de aplicar
mecanismos de control: la primera de ellas es la que llega de las
instrucciones iniciales que se da en una determinada tarea, y que
son las que especifican cuales son las demandas de ésta. La segunda
haría referencia a la actividad que hay que mantener a lo largo de
toda la tarea con el fin de no perder los parámetros de ésta y de no
desviarse de las demandas iniciales. Por última, la tercera señal
vendría del feedback de ejecución, el cual da información de si se
está cumpliendo o no las demandas de la tarea. Dosenbach et al.
(2008) encuentran que el frontal lateral y las regiones parietales se
centran más en el primer tipo de señal, mientras que las zonas
frontales más mediales y el ACC estarían más centrados en el
mantenimiento de las demandas de la tarea.
CAPÍTULO 1
34
Como puede apreciarse, ambos modelos contemplan la
necesidad de que en el cerebro debe existir un mecanismo de
control voluntario que dirija las acciones hacia la consecución de
unos determinados objetivos. Cabe destacar la gran relevancia
posterior que ha tenido el modelo de Norman y Shallice en la
elaboración de teorías posteriores, como la Teoría Atencional de
Posner y Petersen (1990), así como modelos en otros campos, como
puede ser el modelo de la memoria de trabajo de Baddeley. El
modelo del Control Ejecutivo describe cómo ha de funcionar el
sistema de control y como se llevan a cabo diversos procesos
atencionales. Posteriormente, se añaden los mecanismos neurales
que llevan a cabo dichos procesos. Así, tanto el SAS de Norman y
Shallice y la Red Atencional Anterior de Posner comparten tanto
funciones como estructuras cerebrales.
Por otra parte ambos modelos hacen hincapié en que dicho
control se realiza mediante el procesamiento consciente de la
información y de los estímulos, pues solo de esta manera el sistema
sería capaz de ejercer ese control voluntario. En este caso la
consciencia se entiende como un contenido y la definimos como la
capacidad de caer en la cuenta de algo o de percatarnos de lo que
nos rodea. Utilizando la terminología extraída de estudios de lesión
con pacientes humanos, nos estaríamos refiriendo al procesamiento
explícito de la información, en contraposición al procesamiento
implícito, que sería el equivalente al procesamiento no consciente,
más característico de los procesos automáticos. De esta manera
INTRODUCCIÓN
35
podemos decir que todo proceso automático es implícito, aunque no
todo proceso implícito tiene que ser automático.
2. CONTEXTO RECIENTE EN EL QUE SE ENMARCA
NUESTRA INVESTIGACIÓN
En los modelos tradicionales se destaca como punto clave la
existencia de un mecanismo que ante determinadas situaciones
demandantes, va a ejercer su acción de control para ajustar la
ejecución del individuo a las demandas de esta situación, de manera
que esa ejecución sea lo más precisa y ajustada posible.
Sin embargo empieza a surgir la problemática de conocer
realmente, no solo la influencia que el control va a tener sobre la
acción, sino el cómo ese proceso de control se lleva a cabo, de
manera que se pueda tener una teoría del control completa que no
caiga en la tentación explicaciones homunculares, en las que un
determinado mecanismo cerebral “sabe” cuándo intervenir. Es
decir, una buena teoría del control, dad una situación determinada,
debe poder explicar y predecir si va a ser necesaria o no la
aplicación de procesos de control. Es ciertamente comprensible que
no siempre vamos a saber antes de comenzar la realización de una
tarea, que vamos a necesitar del uso de procesos de control. Sin
embargo existen situaciones
A este respecto, Kahneman (1973) argumenta que es el
hecho de intentar realizar una tarea que resulta difícil lo que hace
que se demanden recursos cognitivos. Prueba de esto son los
resultados obtenidos con la tarea Stroop, en la cual los participantes
CAPÍTULO 1
36
muestran mayor interferencia en los dos primeros ensayos de cada
bloque, mientras que dicha interferencia disminuye a medida que se
va avanzando en la serie de ensayos (Henik, Bibi, Yanai y Tzelgov,
1997). Otro de los aspectos que tenemos que tener en cuenta es el
hecho de explicar, una vez que el control cognitivo está siendo
aplicado, cómo se va a modular sus influencia para hacerlo más
eficiente, ya que se ha documentado extensamente que los ajustes
que se hacen en el proceso de control cognitivo, ocurren on-line,
mientras la tarea está siendo llevada a cabo. Así por ejemplo, justo
después de la comisión de un error se observa un aumento, tanto del
tiempo de reacción como de la precisión de respuesta (Laming,
1968; Rabbitt, 1966). Por último, otro punto crucial que debe tener
en cuenta toda teoría del control es conocer aquellos procesos que
van a marcar cuándo y cómo el control cognitivo deja de aplicarse.
Como se ha podido observar, la práctica lleva a la automatización
de las acciones (Anderson, 1982; Shiffring y Schneider, 1977). Este
fenómeno va a hacer que la necesidad de control disminuya, por
tanto debe haber un elemento evaluador que determine cuándo
puede retirarse en control sin que haya un perjuicio en la ejecución
de la tarea.
En resumen, podemos decir que para que el control cognitivo se
lleve a cabo de manera eficaz y eficiente, deben haber dos
dimensiones a tener en cuenta: Una dimensión reguladora que será
la que dé cuenta de la aplicación de procesos top-down, y una
dimensión evaluadora que será la que monitorice el procesamiento
de la información teniendo en cuenta en todo momento las
demandas actuales de la tarea.
INTRODUCCIÓN
37
2.1. La Hipótesis de Monitorización del Conflicto (Botvinick,
Braver, Barch, Carter, & Cohen, 2001).
De acuerdo con lo expuesto, Botvinick, Braver, Barch,
Carter y Cohen (2001) desarrollaron una teoría del control que
intentaba dar cuenta de todos estos aspectos que acabamos de
comentar, en la que se presta especial atención a la dimensión
evaluativa del control cognitivo. Ellos proponen la existencia de un
mecanismo de monitorización del conflicto, que se encargará de
evaluar y monitorizar cuando se produce un conflicto durante el
procesamiento de la información. Su principal función es la de
traducir esa información, en ajustes compensatorios de control. De
esta manera, primero se evalúa la situación y posteriormente se
manda dicha información a aquellos centros responsables de ejercer
los procesos de control, que a su vez se encargarán de proyectar esa
información a los centros necesarios para potenciar su influencia en
el procesamiento.
La estructura que principalmente se encarga de la monitorización
del conflicto es el Cortex Cingulado Anterior (ACC), que detecta y
evalúa la información que puede ser conflictiva. Esta estructura
está conectada con estructuras más prefrontales, como el Cortex
Prefrontal Dorsolateral (DLPFC), a las que manda la información
para que se implementen las estrategias de control propiamente
dichas.
Esta hipótesis se ha formulado como resultado de la múltiple
evidencia de la participación del ACC cuando ocurre una situación
de conflicto, números estudios que han estudiado el papel de dicha
CAPÍTULO 1
38
estructura en la cognición. Estos estudios han sido realizados
utilizando una gran variedad de metodologías, entre las cuales se
incluyen los estudios de lesión, los registros mono-neuronales y las
técnicas de registro de la actividad cerebral, como pueden ser los
potenciales evocados asociados a un evento (ERPs) y en años más
recientes los de resonancia magnética funcional (fMRI), siendo
estos últimos los que mayor cuerpo de datos han arrojado. La
mayoría de los resultados que se han obtenido utilizando estas
metodologías apuntan hacia la participación esencial del ACC en la
detección del conflicto.
A continuación, vamos a realizar un repaso de los
principales datos que existen a este respecto.
2.2 Evidencias a favor de la Teoría de Monitorización del
Conflicto.
En líneas generales, cuando nos enfrentamos a la revisión de
estudios acerca del rol del ACC en la detección del conflicto, nos
encontramos con la posibilidad de clasificar los estudios en tres
categorías diferenciadas: la primera es la que incluye los estudios
que tratan de la necesidad de inhibir una respuesta que es muy
prominente pero totalmente innecesaria para el objetivo de la tarea.
La tarea más frecuente que se ha usado en estos estudios ha sido la
tarea Stroop (Stroop, 1935; McCleod, 1991). Usando esta tarea, el
primero que observó activación en el ACC usando como técnica de
imagen Tomografía por emisión de positrones (PET) fue Pardo
(Pardo, Janer y Raichle, 1990). Lo que observaron fue una mayor
INTRODUCCIÓN
39
activación en el ACC durante la ejecución de la condición
incongruente en comparación con la condición congruente. Este
resultado ha sido replicado en estudios similares, usando fMRI
(Carter, Cohen y Mintun, 1995, Van Veen y Carter, 2006), siendo
también esa activación del ACC en la condición incongruente
mayor que la activación en la condición neutra (Bench et al. 1993;
Carter et al., 1995). Pero no solo se han obtenido estos datos usando
la tarea Stroop, sino también usando otras tareas en las que se
requiere que se inhiba una respuesta automática. Un ejemplo es un
estudio en el que se le pide a los participantes que nombren una
serie de letras B, J, Q e Y. Posteriormente se le pide a esos mismos
participantes que respondan a las letras siguiendo una serie de
reglas simples. Por ejemplo cuando aparezca la J deben responder F
(Taylor, Kornblum, Minoshima, Oliver y Koeppe, 1994). En esta
segunda condición, en la que los participantes deben inhibir la
lectura automática de las letras, se ve una mayor activación del
ACC.
Otro tipo de estudios que pueden enmarcarse dentro de esta
categoría y también han mostrado un incremento en la activación
del ACC, son aquellos que se han realizado usando tareas go-no go.
Casey, Castellanos, F.X., Giedd, Marsh, Hamburguer, Schubert,
Vauss, Vaituzis, Dickstein, Sarfatti y Rapoport, 1997 realizaron
estudios de este tipo usando fMRI. En este estudio los participantes
respondían presionando una tecla cada vez que se presentaba una
letra, excepto cuando esa letra era una X. Para que el hecho de
presionar la tecla se convirtiera en una respuesta prepotente, en la
mayoría de los ensayos no aparecía la X. Cuando comparaban la
CAPÍTULO 1
40
activación del ACC durante esta tarea con la activación del ACC en
una condición control en la cual nunca aparecían las Xs, se observó
una mayor activación en la condición experimental del ACC, sobre
todo cuando aparecía la X. Para generalizar podemos decir que en
todas estas situaciones se da una competición entre vías de
procesamiento que llevan a una respuesta correcta, pero que son
débiles y aquellas vías de procesamiento que llevan a una respuesta
incorrecta y que son prepotentes. Es lo que comúnmente se conoce
como conflicto de crosstalk.
El segundo grupo de estudios acerca de la activación del
ACC y que pueden constituir otra categoría es aquel que hace
referencia a la selección de una respuesta de entre un grupo de
posibilidades entre las cuales no destaca ninguna como más obvia
que las demás para la correcta consecución de la tarea. En este caso,
el estímulo que se presenta, por si solo no da información ni
especifica cual es la respuesta correcta. Un ejemplo de este tipo de
tareas son las tareas denominadas de generar versus repetir.
Pettersen, Fox, Posner, Mintun y Raichle (1988, 1989), llevaron a
cabo una serie de experimentos usando PET en los cuales se les
pedía a los participantes que generaran un verbo que identificara
una acción apropiada a un objeto cuyo nombre se les había
presentado previamente. En otra condición, los participantes
simplemente tenían que leer o repetir el nombre sin generar ningún
verbo. En este tipo de estudios se observa una mayor activación del
ACC en la condición de generar en comparación con la activación
en la condición repetir. Estos estudios han sido replicados por otros
autores (ver Barch, Sabb, Braver y Noll, 2000), incluso si la
INTRODUCCIÓN
41
generación del verbo se hace en silencio (Warburton et al., 1996;
Wise et al., 1991).
Otro ejemplo de este tipo de tareas, son las tareas de fluidez
verbal, en las cuales los participantes deben decir una serie palabras
que empiecen por una letra determinada. En estos estudios también
se observa una mayor activación del ACC cuando los participantes
tienen que realizar esta tarea, en comparación a cuando
simplemente tienen que repetir una palabra que oyen (Frith, Friston,
Liddle y Frackowiak, 1991a) o en comparación a cuando tienen que
realizar una tarea de decisión léxica (Frith, Friston, Liddle y
Frackowiak, 1991b).
También se han realizado estudios con fMRI usando tareas de
fluidez verbal en silencio (Yetkin et al., 1995), así como usando
tareas de fluidez semántica, en los que los participantes deben
generar una serie de nombres correspondientes a una determinada
categoría (Yetkins et al., 1995). Ambos tipos de fluidez, la léxica y
la semántica, han mostrado la producción de una mayor activación
en el ACC.
Las conclusiones que podemos sacar de este tipo de estudios
es que, dado que todo el abanico de respuestas pueden llevar a la
consecución de la tarea, el cerebro activa en paralelo múltiples vías
de procesamiento que son incompatibles entre sí, produciendo lo
que antes hemos llamado un conflicto de crosstalk, esta vez en el
espacio de tiempo entre la presentación del estímulo y la ejecución
de la respuesta.
Por último, vamos a hablar del tercer grupo de estudios que
se han llevado a cabo para comprobar el papel del ACC en el
CAPÍTULO 1
42
proceso de monitorización del conflicto. Estos serían los estudios
que evalúan la comisión de errores. La particularidad que
presentan estos estudios es que en lugar de usas técnicas de imagen
propiamente dichas, usa en su mayoría técnicas electrofisiológicas,
en concreto los potenciales evocados (ERPs) (Rugg y Coles, 1995).
Estos estudias han mostrado que tras la comisión de un error en
tareas de velocidad de respuesta, se observa la aparición de un
potencial negativo, al que se ha llamado Negatividad asociada al
error (ERN) (Falkenstein, Hohnsbein y Hoorman, 1995). Esta
negatividad ocurre 100-150 ms después de que se produzca la
respuesta, e incluso puede producirse antes de que se produzca esa
repuesta errónea, si la persona es consciente de que va a cometer un
error, como se ha observado en estudios en los que se registra este
potencial asociado a una respuesta electromiográfica de la mano
que no corresponde a la respuesta correcta que tiene que dar el
participante (Ghering, Coles, Meyer y Donchin, 1990).
Dehaen, Posner y Tucker, (1994) observaron este potencial
en sus estudios, en los cuales por una parte piden a los participantes
que pulsando una tecla lo más rápidamente posible, digan si un
número (que puede estar representado en su forma numérica o
escrito en palabra) es mayor o menos que 5, y por otra parte les
piden que digan si una serie de palabras que se les presentan son
animales. Cuando los participantes cometían un error y se daban
cuenta de ello, se observaba esta negatividad.
Este negatividad ha sido consistentemente, asociada a regiones
frontales mediales. Mediante el uso de extracción de dipolos se ha
asociado la generación de esta negatividad al ACC (Dehaen et al.
INTRODUCCIÓN
43
1994). Es importante apuntar que junto a los estudios de estimación
de dipolos, de los cuales se han extraído estas conclusiones, se han
realizado estudios usando fMRI con el fin de observar si realmente
es el ACC la estructura implicada en la detección consciente de un
error (Kiehl, Liddle, & Hopfinger, 2000; Menon, Adleman, White,
Glover, & Reiss, 2001). Kiehl, Liddle, & Hopfinger, (2000) usaron
una combinación de ERPs y fMRI con el fin de observar los
correlatos neurales, tanto de respuestas correctas (rechazos
correctos y aciertos) como de las respuestas incorrectas (errores),
durante la realización de una tarea go/no go. Al analizar las
respuestas incorrectas observaron una gran activación del ACC
rostral, así como del cortex frontal lateral izquierdo. Estas áreas solo
se activaron ante los errores y no ante las respuestas correctas, lo
cual sugiere que dichas áreas forman parte de un sistema cerebral
integral de detección del error.
De toda esta revisión de estudios podemos extraer la
conclusión de que el ACC es clave en la detección consciente del
conflicto. Solo cuando el cerebro se percata de que existe una
situación donde existe una competición entre respuestas
automáticas o respuestas que requieren más control, o cuando se
dan varias posibilidades que compiten en igualdad de condiciones o
cuando se ha detectado la comisión de un error, el ACC interviene
monitorizando esa situación de conflicto y mandando señales al
DLPFC que, en el primer caso inhibirá la respuesta automática más
fuerte pero incorrecta y potenciará la respuesta más débil pero a la
vez correcta, mientras que en el segundo caso dará más activación a
una de las respuestas destacándola por encima de las demás, de
CAPÍTULO 1
44
manera que se pueda usar para la tarea, por último en el tercer caso
tras la comisión de un error se pondrán en marcha estrategias de
control para evitar que ese error vuelva a ocurrir, como por ejemplo
aumentar la velocidad de respuesta. Es más, existen estudios con
fMRI, que muestran una doble disociación en la actividad de estas
dos estructuras (McDonald, Cohen, Stenger y Carter, 2000). Estos
autores usaron una variante de la tarea Stroop con el fin de
demostrar la doble disociación entre el ACC, cuya función para
ellos era la de evaluación de la situación y el DLPFC, que para ellos
tenía funciones de mantenimiento de la atención en las demandas de
la tarea. En la tarea daban una instrucción antes de cada ensayo que
bien podía ser leer la palabra (más automático) o bien podía ser
nombra el color de la tinta (se requiere mayor control).
Posteriormente, tras un periodo de tiempo presentaban el target.
Esta estrategia hacía que se separan por un lado los procesos
estratégicos que estarían ligados a las instrucciones (requieren
mantenimiento de la atención en las demandas de la tarea), y por
otro lado aquellos procesos que están ligados a la respuesta (que
serían más evaluativos). Lo que observaron fue una activación en el
DLPFC ligada a las instrucciones, pero solo a las que hacían
referencia a nombrar el color y no a las que se referían a leer la
palabra, mientras que no se observó activación del ACC ligada a las
instrucciones. Esa mayor activación del DLPFC ligada a la
instrucción de nombrar el color, se tradujo en un menor efecto
Stroop ante la presentación de un target incongruente. Por otro lado,
se observó una mayor activación del ACC derecho para aquellos
ensayos incongruentes en comparación con los ensayos
INTRODUCCIÓN
45
congruentes. Esa activación era mayor cuanto mayor era la
interferencia tipo Stroop. Así se pudo ver que el DLPFC izquierdo
se activaba de manera selectiva durante el periodo de preparación,
mientras que el ACC derecho se activaba de manera selectiva
durante el periodo de respuesta, por lo que se puede deducir que
ambas estructuras poseen un rol complementario dentro de una red
neural encargada del control cognitivo.
Uno de los puntos fuertes de la Teoría de Monitorización del
conflicto es que da explicación a múltiples fenómenos que han sido
estudiados en psicología cognitiva. Un claro ejemplo de esos
fenómenos son los efectos de ajustes en el control. En el siguiente
apartado vamos a describir estos efectos y la explicación que da esta
teoría de ellos.
3. EFECTOS DE ADAPTACIÓN AL CONFLICTO
Una prueba más de la existencia de un sistema que actúa
evaluando el conflicto y mandando información a centros que se
encargan de su resolución son las pruebas que arrojan los estudios
donde se han podido observar cambios y fluctuaciones en el control,
asociadas a variaciones en la ejecución o a cambios en las
demandas de la tarea. Estas fluctuaciones se les ha dado el nombre
genérico de Ajustes de Control. A continuación vamos a hacer un
repaso por los efectos de ajuste en el control más comúnmente
observados en la literatura. Posteriormente pasaremos a describir
como la Teoría de la Monitorización del Conflicto, trata de dar
cuenta de cada uno de estos efectos.
CAPÍTULO 1
46
3.1 Efectos secuenciales: Efecto Gratton.
Un efecto secuencial puede definirse como el cambio en el
tiempo de reacción (RT) de la respuesta de un ensayo debido a la
naturaleza del ensayo precedente. Este efecto, ya conocido antes de
que se formulara la hipótesis de monitorización del conflicto,
descrita con anterioridad, fue observado por Cratton, Coles y
Donchin (1992) en un estudio realizado con una modificación de la
tarea de flancos de Eriksen (Eriksen y Eriksen, 1974). En esta tarea
se pide a los participantes que digan, mediante la presión de dos
teclas, la dirección hacia la que apunta una flecha que aparece en el
centro de la pantalla. Esta flecha está flanqueada a derecha e
izquierda por otras dos flechas que pueden apuntar, bien en la
misma dirección que la flecha target (ensayos congruentes), bien en
la dirección contraria (ensayos incongruentes). Si tenemos en
cuenta la secuencia que compondrían dos ensayos se obtienen
cuatro posibilidades: ensayos congruentes que están precedidos por
otro ensayos congruente (CC), ensayos congruentes que están
precedidos por un ensayo incongruente (IC), ensayos incongruentes
que están precedidos por un ensayo congruente (CI) y por último,
ensayos incongruente que están precedidos por otro ensayo
incongruente (II). El efecto que estos autores observaron se produce
de manera más prominente en los ensayos incongruentes, de manera
que ellos compararon la condición CI y la condición II: Lo que se
observa es una reducción del RT en la condición II en comparación
con la condición CI. Asimismo el RT se ve incrementado en la
INTRODUCCIÓN
47
condición IC en comparación con la condición CC. A esta
reducción del RT es a lo que llamaron Efecto Gratton.
La explicación que la teoría de Monitorización del Conflicto
da a este efecto es que en un ensayo incongruente el sistema
necesita recopilar recursos de control para la resolución del
conflicto que estos ensayos general, lo cual prepara al sistema para
resolver futuras situaciones de conflicto. Por tanto, si lo que sigue
es otro ensayo incongruente, el sistema se encontrará altamente
preparado para resolver esa situación de conflicto, lo que se
traducirá en una menor interferencia y en una disminución del RT.
Por el contrario, si el ensayo que sigue es congruente, situación que
se resuelve de manera automática, el sistema se encontrara
demasiado preparado para el control, por lo que volver a una
respuesta automática tendrá un coste y se observará un
enlentecimiento de la respuesta. De la misma manera podemos
explicar el caso de los ensayos congruentes. Como ya hemos
mencionado los ensayos congruentes no suponen ningún tipo de
conflicto al sistema y se resuelven de manera automática. Por tanto,
si un ensayo congruente va seguido de otro ensayo congruente, este
último experimentará una facilitación que hará que el RT aún sea
menor. Si por el contrario va seguido de un ensayo incongruente el
RT aumentará de manera considerable, pues el sistema no está en
absoluto preparado para la resolución del conflicto.
Otra de las predicciones que hace la Teoría de la
Monitorización del conflicto en lo que se refiere a estos efectos, es
que una mayor activación en el ACC se traducirá en un
consiguiente incremento de activación en el DLPFC., que es el
CAPÍTULO 1
48
responsable de implementar las estrategias de resolución del
conflicto y por tanto sería el encargado de realizar los ajustes de
control en función de cada secuencia de ensayos Kerns et al. (2004)
realizaron estudios de fMRI en los que usaban la tarea Stroop.
Eliminaron las repeticiones de estímulos entre ensayos con objeto
de comprobar si se seguían produciendo los efectos secuenciales y
observaron una mayor que una activación del ACC en un ensayo se
traducía en una mayor activación del DLPFC en el siguiente,
obteniendo como predice la hipótesis un menos RT en ensayos II en
comparación con ensayos CI, así como una menor actividad del
ACC en los primeros en comparación con los segundos.. A su vez
observaron que la actividad del DLPFC se incrementaba en aquellos
ensayos en los que mayor ajuste de control se había producido,
confirmando así tanto comportamentalmente como anatómicamente
la Hipótesis de Monitorización del Conflicto.
3.2. Efectos de la frecuencia del tipo de ensayo: Efecto de
proporción de congruencia (PCE), Efecto de compatibilidad
específico del ítem (ISCE) y Efecto de compatibilidad específico del
contexto (CSCE).
Los efectos de ajuste en el control no solo se han observado
al estudiar los efectos secuenciales. Se ha observado que la
frecuencia real o relativa de la incongruencia, produce una
reducción del RT en ensayos incongruentes, sólo cuando dicha
frecuencia es alta.
INTRODUCCIÓN
49
3.2.1. Efecto de proporción de congruencia (PCE)
Cuando hablamos de este efecto, nos referimos a una
disminución de la interferencia en ensayos que implican conflicto,
cuando la proporción de ensayos incongruentes es elevada en
comparación con cuando esta proporción de ensayos incongruentes
es baja. Asimismo, los ensayos no conflictivos se procesan más
rápido cuando los ensayos incongruentes son infrecuentes en
comparación con cuando la proporción de ensayos incongruentes es
alta (Logan y Zbrodoff, 1979; Logan et al., 1984).
Este efecto se ha podido observar en numerosos estudios que
usan la tarea Stroop. Logan y Zbrodoff (1979) llevaron a cabo un
estudio en el que usaron una variante de la tarea Stroo en la que los
participantes tenían que decir la posición de una palabra (ARRIBA
y ABAJO), que podía presentarse por encima o por debajo de un
punto de fijación. O bien tenían que decir el significado de esas dos
palabras. En ambas tareas los ensayos podían ser compatibles
(ARRIBA presentado por encima del punto de fijación, ABAJO
presentado por debajo del mismo) o bien podían ser incompatibles
(ARRIBA presentado por debajo del punto de fijación y ABAJO
presentado por encima). Estos autores variaron la proporción de
ensayos compatibles e incompatibles (20% de incompatibles para la
condición de baja incongruencia y 80% de incompatibles para la
condición de alta). Lo que observaron fue que en la tarea donde
tenían que decir el significado de la palabra, cuando los
participantes estaban en la condición de baja incongruencia
procesaban de manera más rápida los ensayos compatibles, mientras
CAPÍTULO 1
50
que cuando estaban en la condición de alta, lo que obtenían la
facilitación eran los ensayos incompatibles. Otros estudios
confirman estos resultados usando la tarea Stroop o variantes
(Cheesman y Merikle, 1986; Lindsay y Jacoby, 1994; Logan, 1980;
West y Baylis, 1998; Kane y Engle, 2003 experimento 4). En todos
estos estudios la manipulación que se realiza de la proporción de
congruencia es a un nivel global de lista, donde se varía el número
total de ensayos incongruentes y de ensayos congruente.
La explicación que da la Teoría de la monitorización del
conflicto sobre este efecto es que ante una situación en la que la
incongruencia y por tanto el conflicto es frecuente, hace que el
sistema de control esté mejor preparado para enfrentarse a futuras
situaciones de conflicto, por lo que cuando se presente un ensayo
incongruente se procesará más rápidamente y habrá una menos
diferencia en lo que se refiere a RT entre ensayos incongruentes y
congruentes. Los participantes generarían expectativas acerca de
qué tipo de ensayo va a ocurrir a continuación, desarrollando
estrategias de control de manera proactiva (Braver et al., 2007).
Cuando un participante espera que aparezca un ensayo congruente,
como ocurriría en aquellos bloques de ensayos donde los
proporción de congruentes es mayor, desarrollarán la estrategia de
leer la palabra. Esta estrategia hará que se facilite el procesamiento
de los ensayos congruentes, empeorando por tanto el procesamiento
de los incongruentes, lo que se traduciría en mayor efecto Stroop.
Por el contrario, cuando el participante se encuentra en un bloque
donde la proporción de ensayos incongruentes es mayor, hará un
doble esfuerzo por evitar el procesamiento del significado de la
INTRODUCCIÓN
51
palabra, lo que hará que empeore su percepción de ensayos
congruentes, pero mejorará sustancialmente la percepción de
aquellos ensayos que sean incongruentes, disminuyendo
significativamente la interferencia de tipo Stroop. La base neural de
esta explicación es que el ACC detectaría esa situación de conflicto
frecuente, de modo que el DLPFC estaría altamente preparado para
implementar la resolución de ese conflicto. Este sistema, al
encontrarse activo y facilitado, estaría mejor preparado para
resolver una futura situación de conflicto y por tanto se ve esa
disminución de la interferencia.
3.2.2. Efecto de proporción de congruencia específico del ítem
(ISPCE).
Esiste un grupo de autores, sin embargo que proponen que el
efecto de proporción de congruencia no ocurre a nivel de lista, sino
que actúa a nivel de ítem propiamente dicho. De acuerdo con estos
autores la adaptación al conflicto no necesita de una gran secuencia
de ensayos o de una manipulación a nivel de bloque para poder
observare, si no que dichos ajustes se van produciendo ya ensayo a
ensayo. Jacoby, Lindsay & Hessels (2003) obtuvieron este efecto
usando de Nuevo la tarea Stroop. Eligieron un conjunto de seis
colores y sus correspondientes 6 nombres de color. Para generar la
condición de ítems mayoritariamente congruentes (MC) los
nombres de palabras de uno de los conjuntos fueron presentadas en
su color correspondiente un 80% de las veces y en otro color
diferente un 20%. Las palabras del otro conjunto revirtieron este
CAPÍTULO 1
52
orden con el fin de producir ítems mayoritariamente incongruentes
(MI). La proporción de congruencia global se mantenía al 50% de
congruentes y 50% de incongruentes, lo cual evita que los
participantes desarrollen estrategias generales para adaptarse al
conflicto. Lo que Jacoby et al. encuentran es que los ítems MC
mostraban un mayor efecto Stroop que los MI. La proporción
relativa de incongruencia que se asociaba a cada uno de los ítems
ajustaba el sistema, produciendo un mayor efecto de control en
aquellos ensayos incongruentes en los que el ítem era MI.
Según la Teoría de monitorización del conflicto el cerebro
debe detectar que hay más ensayos incongruentes que congruentes
para poder desarrollar una estrategia de control acorde a esa
situación y así resolver mejor una nueva incongruencia. Sin
embargo, en los estudios de Jacoby et al. el cerebro no puede
generar una estrategia global en relación con el porcentaje de
congruencia, pues a nivel global de tarea el porcentaje de
congruentes e incongruentes es el mismo. Es necesario una
actualización on-line de la situación para que se pueda dar ese
beneficio con una proporción de incongruencia que está asociado a
un ítem. Para estos autores el ISPCE estaría reflejando un control
rápido dirigido por el estímulo. Una especie de “control
automático”. Según esta perspectiva cada ítem individualmente se
asocia con el filtro atencional que más frecuentemente se debe
emplear para ese tipo de ítem a lo largo de toda la sesión
experimental. De esta manera, por ejemplo los ítems MC quedarían
asociados con un filtro atencional que filtre de manera muy débil la
lectura de la palabra, mientras que los ítems MI quedarían asociados
INTRODUCCIÓN
53
a un filtro que filtre de manera fuerte la información de la palabra
(Jacoby et al., 1999). Por tanto, cuando un determinado ítem
aparece en la pantalla, inmediatamente de forma reflja activará el
filtro atencional que se haya asociado al tipo de ítem que es y ese
filtro rápidamente ajusta los mecanismos atencionales para
proporcionar control on-line sobre el procesamiento de dicho ítem
Stroop. En conclusión, en lo que se refiere a la tarea Stroop usada
por Jacoby et al. (2003), la influencia de la palabra actuaría a un
nivel de ítem, siendo ese ítem el que actúa de clave contextual para
selecciona un set atencional u otro en función de las demandas de la
tarea.
Existe una visión alternativa que explicaría este efecto como
el resultado de un aprendizaje asociativo específico para cada ítem
que se centra en la frecuencia con la que una palabra determinada y
un determinado color se asocian a lo largo de un diseño
experimental (Jacoby et al., 2003). De ese modo los participantes
aprenden a responder más rápido a aquellos pares de palabra-color
que son más frecuentes en comparación con aquellos pares que no
lo son tanto (Logan, 1988). Los ensayos MC repetirán de forma más
frecuente ítems específicamente congruentes y de manera
infrecuente ítems específicamente incongruentes. Al contrario
sucedería con los MI, por lo que los participantes estarían
respondiendo a la asociación más frecuente para cada ítem
(Schmindt y Besner, 2007 y 2008; Atalay y Misirlisoy, 2012).
Con el fin de tener en cuenta estos resultados y tomar en
cuenta el control automático, Blais, Robidoux, Risko, & Besner
(2007) desarrollaron un modelo computacional que modifica el
CAPÍTULO 1
54
modelo de Botvinick et al. En este modelo, el DLPFC ejercería los
procesos de control a nivel de un ítem particular, en vez de a un
nivel de lista más general. Posteriormente Blais y Bunge (2010)
realizaron un estudio en el que usaron un diseño casi idéntico al del
estudio de Bugg, Jacoby y Toth (2008), mientras registraban la
actividad cerebral con fMRI. A parte de replicar los efectos
comportamentales aportando de nuevo evidencia a favor del control
cognitivo a nivel de ítem, obtuvieron que el ACC y el DLPFC, que
son estructuras tradicionalmente asociadas al control de tipo top-
down (Botvinick et al., 2001), se activaban de manera selectiva ante
condiciones en las que actuaban procesos de control específicos del
ítem. Sin embargo, dicha activación no se daba cuando la
proporción de congruencia relativa al ítem era del 50%, lo cual
aporta evidencias a favor de la existencia de mecanismos de control
automáticos, que están siendo llevados a cabo por estructuras
frontales que se han asociado a control consciente y voluntario.
3.2.3. Context-Specific Compatibility Effect
La tercera forma que se usa para manipular el porcentaje de
congruencia y que también ha contribuido a aclarar la distinción
entre control automático y aprendizaje asociativo es la del contexto.
En este caso el porcentaje de congruencia se varía en diferentes
contextos en los cuales se presentan los mismos ítems. La lógica
que sigue esta manipulación es la siguiente: si como ocurre en el
ISPCE un ítem puede por sí mismo servir de estímulo para alicitar
procesos de control automáticos on-line que inciden directamente
INTRODUCCIÓN
55
sobre los filtros atencionales a usar, según las demandas de la tarea,
entonces un determinado contexto que se asocia a un determinado
ítem, como por ejemplo puede ser la localización espacial en la que
aparece, podrá servir de señal también a la hora de ajustar los
procesos rápidamente seleccionando el filtro atencional que
corresponda a las demandas actuales de la tarea. Crump et al.
(2006) diseñaron un estudio en el que utilizaban una versión de la
tarea Stroop en la que se presentaba una palabra que iba seguida de
un parche de color que podía presentarse o bien por encima de un
punto de fijación o bien por debajo de este. La tarea de los
participantes era responder al parche de color, que podía ser
congruente o incongruente con respecto a la palabra (prime). La
localización del parche de color (arriba o abajo del punto de
fijación) era lo que definía el contexto para la manipulación del
porcentaje de ensayos congruentes. Así aquellos parches de color
que aparecían arriba eran asociados con un porcentaje de
congruencia alto (75% de ensayos congruentes), mientras que
aquellos que aparecían por debajo estaban asociados a un porcentaje
de incongruencia alto (75% de incongruentes). En este tipo de
manipulaciones, el porcentaje de congruencia a nivel de lista se
mantiene al 50/50. Lo que se obtuvo fue un menor efecto Stroop
cuando el contexto era mayoritariamente incongruente (MI),
mientras que dicho efecto Stroop se incrementaba en el contexto
mayoritariamente congruente (MC). Otros estudios usan como
contexto el tipo de letra en el cual aparecen escritas las palabras, en
lugar de utilizar la localización espacial (Bugg et al., 2008)
obteniendo los mismos resultados.
CAPÍTULO 1
56
De nuevo aquí surge la duda de si este efecto es debido a
procesos de control automático o a los efectos de aprendizaje
debido a la frecuencia de un tipo u otro de ensayos o ítems. Para
intentar desentramar esta cuestión Crump y Miliken (2009)
realizaron un estudio en el que manipularon el porcentaje de
congruencia, tanto a nivel de contexto como a nivel de ítem. En el
primer experimento usaron la misma variante de tarea Stroop en la
que se presentaban palabras (primes) y parches de color (targets).
En este diseño había dos localizaciones (arriba y abajo del punto de
fijación) que eran 100% predictivas solo para un subconjunto de
palabras-parches de color. A este subconjunto lo llamaron context
probes. Por lo tanto, el contexto localización no era predictivo para
el resto de ítems. A estos los llamo transfer ítems. Esta
manipulación evitaba que los transfer probes se beneficiaran de la
variable frecuencia, pues cada transfer ítem aparecía el mismo
número de veces. La hipótesis que tenían estos autores es que si el
CSCE solo se debe a procesos de aprendizaje que se desarrollan por
la presentación de un ítem de manera frecuente, no podremos
observarlo en los transfer probes, si no solo en los context probes.
Por el contrario, si el CSCE se debe a modulaciónes que proceden
de las características del contexto, sí que se observará este efecto,
tanto en los context probes como en los transfer probes. Es más. Lo
participantes aprenderán a aplicar determinados parámetros de
atención selectiva para controlar la ejecución de la tarea en los
context probes, y transferirán ese aprendizaje a la ejecución en los
transfer probes. Este efecto de transferencia que requiere de la
experiencia de los participantes con los context probes se potenciará
INTRODUCCIÓN
57
con la práctica y será menos visible al principio de la sesión.
Efectivamente observaron CSCE en los transfer probes. Pese a que
estos ítems se presentaban con la misma frecuencia, sufrieron las
influencias del contexto, mostrando un mayor efecto Stroop cuando
se presentaban en el contexto MC en comparación con los que
presentaban en el contexto MI. Este efecto de transferencia supone
evidencia clara la existencia de ajustes en el control que se
producen de manera rápida, on-line y automáticamente generados
por el contexto, puesto que los procesos que se han ido aplicando
para los context probes, en una situación totalmente predictiva, se
han trasferido de manera automática a los transfer probes, que se
presentan en una situación en la que los participantes no pueden
predecir ni anticipar el tipo de estrategia que deben utilizar.
Aunque por motivos de interés en esta tesis nos estamos
basando en los estudios que utilizan la tarea Stroop o alguna
variante de esta, similares resultados en la manipulación del
porcentaje de congruencia se han obtenido con otro tipo de
paradigmas, como por ejemplo la Tarea de Flancos (Eriksen y
Eriksen, 1974), tanto en las manipulaciones de este porcentaje a
nivel de lista, como de ítem o de contesto (Corballis y Gratton,
2003; Lehle y Hübner, 2008).
4. ACLARANDO EL CONCEPTO DE CONSCIENCIA
Como hemos visto al estudiar los fenómenos de adaptación
al conflicto, el cerebro es capaz de procesar ciertos aspectos de la
información a unos niveles complejos sin que la persona sea
CAPÍTULO 1
58
consciente de dicho procesamiento, lo que nos lleva a plantearnos
más específicamente la cuestión de la existencia de procesos de
control no conscientes. El interés por el estudio de dichos procesos
se remonta a los comienzos de la Psicología Experimental. Una de
las primeras estrategias que se usaron para estudiar estos procesos
experimentalmente, fue la degradación de un estímulo hasta que la
persona era incapaz de percibirlo. Otra de las estrategias era
disminuir cada vez más el tiempo de presentación de dicho
estímulo, de modo que llegara un momento en el que se presentaba
lo suficientemente rápido para que la persona no pudiera percibirlo.
Al valor crítico de intensidad o de tiempo de exposición de un
estímulo por debajo del cual una persona es incapaz de percibir
dicho estímulo, se le llamó umbral de consciencia. Este umbral
puede ser de dos tipos: podemos hablar de umbral subjetivo de
consciencia, cuando nos referimos a la condición mínima de
estimulación necesaria para que una persona tenga experiencia de
percibir un determinado estímulo. De la misma manera hablamos
de umbral objetivo de consciencia cuando nos referimos a la
condición mínima de estimulación por debajo de la cual una
persona ejecuta una tarea al azar.
Podría decirse que las medidas de umbral objetivo y umbral
subjetivo forman parte de un continuo de percepción, en el que el
umbral objetivo marcaría el punto en el que tenemos la total certeza
de que una persona no es consciente de lo que se le está presentando
y donde el umbral subjetivo sería el punto en el que estando más
próximos a la percepción consciente, la persona siente que no es
capaz de percibir aquello que se le está presentando. Cabe resaltar la
INTRODUCCIÓN
59
importancia que en investigación tiene el umbral subjetivo de
consciencia, pues es el que define el término de consciencia
propiamente dicho. Aquellos estímulos que se sitúan por debajo del
umbral subjetivo no son reportables por la persona, como ocurre en
determinadas patologías neuropsicológicas como el neglect o el
blindsight. Sin embargo, aunque no existe reportabilidad, la
ejecución de pacientes con estos déficits en tareas de detección está
por encima del azar.
A nivel teórico, Dehaen y Changeoux (2005) desarrollaron
un modelo neural de espacio de trabajo en el que postulan que las
proyecciones colinérgicas que parten del bulbo raquídeo en sentido
ascendente, mandan señales neuro-moduladoras al tálamo y la
corteza. Para que esas señales puedan ser enviadas, debe existir un
estado de vigilancia óptimo. Se ha observado que un buen estado de
vigilancia correlaciona con actividades en el bulbo raquídeo y en el
tálamo y posteriormente en el córtex, cobrando importancia las
conexiones funcionales entre el cortex prefrontal y ACC (Balkin, et
al., 2002). Como vemos estas estructuras corticales son las mismas
que se asocian al control cognitivo, por lo que esto apoyaría la
necesidad de la consciencia para que estos procesos puedan llevarse
a cabo.
En base a este modelo teórico Dehaene, Changeux,
Naccache, Sackur y Sergent (2006) desarrollaron una nueva
tipología conceptual para clasificar y diferenciar diferentes tipos de
procesamiento estimular. Según ellos habría tres tipos: consciente
sería aquel tipo de procesamiento que permite que la persona pueda
reportar lo que está percibiendo. Para que se de este tipo de
CAPÍTULO 1
60
procesamiento es necesario y fundamental que se activen las
proyecciones frontales, que comunican las estructuras subcorticales
con la corteza cerebral.
Por otra parte pre-consciente sería aquello que aún no
llegando en un principio a ser consciente, podría serlo si se dirige la
atención hacia el estímulo. Numerosos ejemplos de procesamiento
pre-consciente se ha encontrado en la literatura y en la historia de la
Psicología. De esta manera, es en esto en lo que se basaba la terapia
psicológica de Freud, que pretendía que la persona atendiera a
aspectos de su psique que se mantienen a un nivel inconsciente,
pero que con las técnicas adecuadas puede salir a la luz y ser
tratado. Otros ejemplos son el parpadeo atencional, la ceguera por
inatención o la ceguera al cambio.
Por último ellos se utilizaban el término subliminal para
hacer referencia a aquello que aunque la persona le preste atención
la información no llega a acceder a la consciencia, porque de alguna
manera el procesamiento queda interrumpido y no puede darse la
reverberación de las proyecciones frontales. Un claro ejemplo de
este fenómeno sería el proceso de enmascaramiento. En dicho
proceso el estímulo se presenta o bien demasiado rápido como para
que la persona no sea capaz de procesarlo, o bien degradado hasta
un punto en el que el procesamiento no es posible o bien seguido de
otro estímulo que actúa de máscara y que impide que el
procesamiento del primer estímulo pueda llevarse a cabo. Este tipo
de experimentos se han realizado manipulando también el
porcentaje de congruencia, a fin de observar si se producen efectos
de adaptación al conflicto a nivel no consciente. De esta manera, los
INTRODUCCIÓN
61
participantes no pueden ser conscientes de la relación de conflicto
existente entre el prime y el target y por tanto no pueden establecer
ninguna estrategia deliberada para adaptarse al conflicto (Klapp,
2007 experimentos 2 y 3). Lo que estos estudios muestran es que
incluso así, los participantes presentan efectos de adaptación al
conflicto. No se puede equiparar el término consciencia con el
término percepción, de manera que ser consciente de algo es
percibir ese algo, puesto que la percepción incluye multitud de
procesos de los que la persona no es consciente. El resultado de
percibir requiere de una serie de procesos previos que no van a
acceder a la consciencia, de modo que sin ellos no habrá
percepción.
Una vez que se han estimado los valores de umbral
perceptivo, se usan dichos valores como tiempo de presentación de
los estímulos de la fase experimental.. Numerosos estudios
muestran que el valor crítico de tiempo de presentación de un
estímulo, por debajo del cual éste ya no es conscientemente
percibido se sitúan en torno a los 30 ms (Greenwald, Klinger, &
Liu, 1989; Daza et al., 2002; Klapp, 2007, Ortells, Marí-Beffa, &
Plaza-Ayllón, 2012).
Mediante el uso de este tipo de experimentos, usando
valores tanto de umbral objetivo como de umbral subjetivo, se ha
podido comprobar que las personas son capaces de discriminar
numerosos aspectos de los estímulos de manera no consciente,
llegando incluso a ser capaces de discriminar el significado de las
palabras (Ruz, Madrid, Lupiáñez y Tudela, 2003) y la expresión
emocional de las caras.
CAPÍTULO 1
62
5. NUESTRA INVESTIGACIÓN
El objetivo principal de nuestra investigación es aportar
evidencia a favor de la existencia, de procesos de control no
conscientes. Con este fin realizamos dos series experimentales en
las que se registró la actividad cerebral usando potenciales
corticales evocados (ERPs), con objeto de estudiar la temporalidad
de los procesos. Consideramos muy importante el uso de este tipo
de técnicas para estudiar los procesos de adaptación al conflicto de
manera no consciente, pues nos pueden dar información sobre los
mecanismos cerebrales que van a estar actuando a la hora de
resolver ese tipo de conflicto, así como el momento en el que dicho
proceso está ocurriendo, cosa que hasta ahora no ha sido descrita en
la literatura sobre esta temática. Nuestro objetivo a este respecto es
ver si realmente se producen efectos de adaptación al conflicto no
consciente, y si se producen, ver si dicha adaptación implica los
mismos mecanismos cerebrales, o similares, a los que están
implicados en el control consciente. Por tanto, este tipo de medidas
proporcionan información adicional sobre cómo el cerebro está
respondiendo ante los estímulos previos (primes), que en nuestro
caso estarán enmascarados, así como sobre la respuesta hacia el
objetivo. También esperamos obtener información acerca de la
relación de congruencia entre ambos, primes y objetivo, y del
procesamiento de la manipulación de la proporción de la
congruencia. (Desender y Van den Bussche, 2012).
La primera serie experimental consta de un experimento en
el cual usamos una variante de la tarea Stroop y seguimos un
INTRODUCCIÓN
63
paradigma de enmascaramiento, mientras se registraba la actividad
cerebral. El objetivo de este estudio fue doble: por una parte
quisimos estudiar el efecto Stroop en condiciones de
enmascaramiento y ver cómo influye la manipulación de la
proporción de congruencia cuando ésta se asocia a un determinado
contexto, que en nuestro caso fueron dos tipos de máscaras
utilizadas para evitar que se produjera el procesamiento consciente
de los primes. Por otro lado quisimos observar el curso temporal de
activación neural que se producía al procesar la proporción de
congruencia y la relación de congruencia entre la palabra que
designa color y el objetivo.
Usamos un conjunto de ensayos en los cuales enmascaramos
una palabra que designaba un color. Previamente creamos una
expectativa sobre la aparición de dicha palabra calibrando de
manera individual el umbral subjetivo por debajo del cual el
participante informaba no percibir dicha palabra, partiendo de una
condición en que la misma era plenamente visible. Así pretendimos
mantener constante la expectativa del prime pese a que éste
permanecía enmascarado. Se ha demostrado que mantener dicha
expectativa puede generar un priming semántico subliminal entre el
prime y el objetivo (Martens, Ansorge & Kiefer, 2011).
Manipulamos, a su vez, el porcentaje de congruencia a nivel de
ítem, manteniendo la proporción de ensayos congruentes e
incongruentes a nivel de lista al 50/50 y asociamos las distintas
proporciones de congruencia con cada uno de los tipos de máscara,
de manera que sistemáticamente una máscara creaba un contexto de
más incongruencia que la otra.
CAPÍTULO 1
64
En la segunda serie experimental nos propusimos comparar
el conflicto cognitivo con el emocional, manipulando las principales
variables relacionadas tanto con el conflicto (congruencia) como
con la adaptación al conflicto (congruencia previa y porcentaje de
congruencia). Para este fin utilizamos la tarea de categorización de
caras (Egner y Hirsch, 2005b), en la cual los participantes debía
clasificar una serie de caras en función del género (tarea no
emocional) o en función de la expresión emocional de las caras
(tarea emocional). En el primer experimento, superpuesta a la cara,
aparecía una palabra que podría ser congruente o incongruente con
el género (tarea no emocional) o con la expresión de las caras (tarea
emocional). En este experimento los participantes tuvieron acceso
consciente a los estímulos, tanto palabras como caras. En el
segundo experimento las palabras aparecieron enmascaradas
previamente a la presentación de las caras, replicando en gran parte
el procedimiento del experimento de la primera serie.
Capítulo 2
SERIE EXPERIMENTAL I:
Unconscious context-specific proportion
congruency effect in a stroop-like task
El contenido de este capítulo ha sido publicado como Panadero, A.,
Castellanos, M. C., & Tudela, P. (2015). Unconscious context-
specific proportion congruency effect in a stroop-like task.
Consciousness and cognition, 31, 35-45.
Abstract
Cognitive control is a central topic of interest in
psychology and cognitive neuroscience and has traditionally
been associated with consciousness. However, recent research
suggests that cognitive control may be unconscious in
character. The main purpose of our study was to further
explore this area of research focusing on the possibly
unconscious nature of the conflict adaptation effect, specifically
the context-specific proportion congruency effect (CSPCE), by
using a masked Stroop-like task where the proportion of
congruency was associated to various masks. We used
electrophysiological measures to analyze the neural cor- relates
of the CSPCE. Results showed evidence of an unconscious
CSPCE in reaction times (RTs) and the N2 and P3 components.
In addition, the P2 component evoked by both target and masks
indicated that the proportion of congruency was processed earlier
than the con- gruency between the color word and the ink color
of the target. Taken together, our results provided evidence
pointing to an unconscious CSPCE.
SE I: EXPERIMENTO 1
68
1. Introduction
In a Stroop task, participants have to name the ink color of
color words (Stroop, 1935, experiment 2). In a more recent
version of the task, two different conditions are usually presented:
congruent and incongruent. In the congruent condition, the ink
color matches the color name of the word (e.g., RED in red ink).
In the incongruent condition, by contrast, the ink color is different
from the color name of the word (e.g., BLUE in red ink). To
perform the task in the incongruent condition, participants must
avoid the automatic process of reading the word and name the
color of the ink in which the word is printed by using control
strategies. In the congruent condition, the ink color matches the
color name of the word, so avoidance of the automatic reading
process is not necessary. Consequently, reaction times (RTs) are
longer in the incongruent condition than in the congruent one. This
additional time is known as Stroop interference or Stroop effect.
Stroop interference is affected by the relative proportion of
congruent and incongruent trials. Behavioral studies have shown a
reduction in RTs in incongruent trials presented in a context of low
congruency (Logan & Zbrodoff, 1979). In other words, when the
proportion of incongruent trials is higher than that of congruent
trials, the Stroop effect is smaller than in situations in which the
proportion of congruent trials is predominant. It seems as if
frequent experience with conflicting stimuli or response features
facilitates the resolution of interference. In general terms, this
UNCONSCIOUS CSPCE IN A STROOP LIKE TASK
69
facilitation is known as the conflict adaptation effect (Botvinick,
Braver, Barch, Carter, & Cohen, 2001).
Over the years, this improvement in control has been
attributed to different factors. The main ones are the overall pro-
portion of congruency at the list level (proportion of congruency
effect, PCE) and a more specific and online processing of the
information either at the item level (item-specific proportion
congruency effect, ISPCE) or at the context level (context-specific
proportion congruency effect, CSPCE). Proponents of the PCE
(Logan, Zbrodoff, & Williamson, 1984) argue that in a Stroop task
subjects become aware of the contingency between the color and
the word and adapt their strategies to attend to the word when the
proportion of congruent trials is higher than that of incongruent
ones and ignore the word when the proportion of congruent and
incongruent trials is reversed. Consequently, the Stroop effect is
larger when the proportion of congruent trials is predominant
because the color word is likely to be attended to in incongruent
trials. By contrast, when the proportion of congruent trials is lower
than that of incongruent trials, the Stroop effect is smaller than in
the previous case because the color word is mostly ignored in
incongruent trials. According to this approach, the proportion of
congruent items influences performance at a list-wide level by
switching participants’ attention to one of the two dimensions
depending on which one is relevant for the task.
At computational and neural levels of analysis, Botvinick et
al. (2001; see also Botvinick, Cohen, & Carter, 2004; McDonald,
SE I: EXPERIMENTO 1
70
Cohen, Strenger, & Carter, 2000) implemented these ideas in what
they termed the conflict-monitoring hypothesis. According to these
authors, the Stroop task is just a particular case of conflict resolved
by the interplay of a set of neural structures. The dorsal anterior
cingulate cortex (dACC) acts as a detection mechanism that
responds to the occurrence of a conflict situation (i.e., an
incongruent trial in a Stroop task). The conflict signal triggers
strategic adjustments in the dorsolateral prefrontal cor- tex
(DLPFC) and other posterior and subcortical structures that serve to
prevent conflict in subsequent performance. Accord- ingly, in a list
where incongruent trials outnumber congruent ones, strategic
adjustment mechanisms will be highly operative, thus minimizing
the influence of the color word on performance and reducing the
Stroop effect. For the purposes of the pres- ent study, it should be
underlined that PCE proponents understand control as a strategic
and conscious process.
Proponents of the ISPCE contend that the proportion
congruency effect acts at the item level. Jacoby, Lindsay, and
Hessels (2003) chose six colors and their corresponding color
words and divided them into two sets of equal size. To produce
mostly congruent (MC) items, words in one set were presented in
their congruent colors in eighty percent of trials and in another
color from that set in the remaining twenty percent of trials. These
rates were reversed in words from the other color set to produce
mostly incongruent (MI) items. The proportion of congruency at the
list level was fifty percent. This manipulation of proportions
prevents subjects from developing a general list-level strategy
UNCONSCIOUS CSPCE IN A STROOP LIKE TASK
71
because the proportion of congruent and incon- gruent items is the
same. However, Jacoby et al. found that MC items showed a larger
Stroop effect than MI items. From a computational approach,
Blais, Robidoux, Risko, and Besner (2007) showed that
Botvinick’s conflict-monitoring model could not explain the
ISPCE and proposed a modified model according to which the
DLPFC exerts control at the specific item level rather than at the
general list level. Results similar to those of Jacoby et al. were
obtained by Crump, Gong, and Milliken (2006) in a situation in
which the likelihood of congruency was associated to a task-
irrelevant location context rather than to a particular color word.
According to these authors, the ISPCE belongs to a larger class of
effects known as context-specific proportion congruency effect
(CSPCE) that is driven by the relationship between the context and
the likelihood of congruency. In recent years, the CSPCE has been
reported in a range of Stroop-like effects (Bugg, Jacoby, & Toth,
2008; Crump, Vaquero, & Milliken, 2008; for a review, see Bugg &
Crump, 2012) and in different types of contexts, including social
categories (Cañadas, Rodríguez-Bailón, Milliken, & Lupiáñez,
2012). An interesting characteristic of the CSPCE is that the type of
control involved cannot be assumed to be a strategic, deliberate
and conscious process. Instead, this type of control seems to be
automatic (Jacoby et al., 2003) and unconscious. In fact, the effect
has been shown to be independent of participants’ awareness of
the proportion of congruency manipulation (Crump et al., 2008).
The main purpose of the present research was to further
explore the likely unconscious character of the conflict adaptation
SE I: EXPERIMENTO 1
72
effect in a Stroop task by asking subjects to name the color of a
color patch that was preceded by a masked color word. We used
two masks (one previous to the color word and one following the
color word) and calibrated the duration of the color word
individually for each participant until it was unnoticed. Under
similar masking conditions to ours, this variable has been used in
some studies in which the proportion of congruency was
manipulated in an effort to find qualitative differences between
conscious and unconscious processes (Daza, Ortells, & Fox,
2002; Merikle & Joordens, 1997). In those experiments, the usual
Stroop effect was observed under masked color word conditions.
Yet, the effect was reversed when the color word was easily seen,
as participants exhibited faster RTs in incongruent trials than in
congruent ones. The PCE effect has also been explored under
masked conditions in other experiments. Klapp (2007, Experiment
3) used a spatial Stroop-like task involving successively presented
arrows pointing in either the same direction (i.e., compatible
condition) or the opposite direction (i.e., incompatible condition).
The first arrow was presented for 32 ms and was immediately fol-
lowed by the mask and the second arrow, so that the first arrow was
unnoticed to the subjects. Results revealed a significant interaction.
Specifically, the difference in RTs between compatible and
incompatible conditions was greater when the arrows pointed in
the same direction than when they pointed in opposite directions.
However, in Klapp’s experiment (2007), participants received
feedback for incorrect responses, which makes it difficult to
interpret his results. A possible explanation could be that subjects
UNCONSCIOUS CSPCE IN A STROOP LIKE TASK
73
adapted their responses to conscious error rates rather than to the
unconscious frequency of congruent and incongruent trials.
Heinemann, Kunde, and Kiesel (2009) explored the CSPCE under
masked conditions using a task in which subjects had to categorize
target numbers as being larger or smaller than five. The target was
preceded by a masked number (i.e., the prime) that could be
congruent or incongruent with the target. At the beginning of each
trial, a colored rectangle was presented as background
simultaneously with the fixation cross. The color of the rectangle
was associated with a particular congruency context that could be
either 80% or 20%. Results showed that the difference in RTs
between congruent and incongruent trials was significantly higher
in the 80% than in the 20% congruent trial condition but only when
subjects were able to see the prime; no differences were observed
when they could not see it. (Heinemann et al., 2009). The authors
concluded that the CSPCE requires conscious representation of the
conflicting information, namely prime, target and context (Kunde,
Reuss, & Kiessel, 2012).
It seems that the unconscious conflict adaptation effect has
not been clearly established. First, studies of the PCE under
unconscious conflict conditions are difficult to interpret due to the
presence of confounding factors. Second, the CSPCE seems to
occur only when the conflicting information is consciously
perceived. We considered that it would be interesting to
explore the conflict adaptation effect by associating the
proportion of congruent and incongruent trials to different masks
and using electrophysiological recordings at the same time to
SE I: EXPERIMENTO 1
74
study the brain response to both conflict adaptation and the
proportion of congruency.
We partly followed a procedure developed by Blais, Tudela,
and Bunge (2007) and used two different types of masks. One type
of mask was associated with mostly congruent trials and a second
type of mask was associated with mostly incongruent trials. Each
mask appeared an equal number of times within each block. Yet,
one mask was followed by 33% congruent and 66% incongruent
color words the other mask was followed by the opposite
proportion of congruent and incongruent trials. Thus, the total
number of congruent and incongruent trials was kept the same in
each block.
The use of brain activity-related measures to study the
unconscious conflict adaptation effect has been highly recom-
mended (Desender & Van den Bussche, 2012). Such measures can
provide additional information about the brain response not only to
targets and masked primes but also to the processing of the
congruency relation between prime and target and of the proportion
of congruency. In an influential study, Dehaene et al. (2003),
Dehaene, Sergent, and Changeux (2003) used functional magnetic
resonance imaging (fMRI) and a Stroop-like task in which subjects
were asked to decide whether target numbers were larger or
smaller than five. Targets were always preceded by another
number (i.e., the prime) that could be congruent with the target
(i.e., both numbers greater or smaller than five) or incongruent
(i.e., one number larger and the other one smaller than five). In
UNCONSCIOUS CSPCE IN A STROOP LIKE TASK
75
addition, the numbers could be either visible or made invisible by
masking. They found that normal subjects showed a clear
activation of the anterior cingulate cortex (ACC) when the prime
could be consciously pro- cessed; by contrast, no ACC activation
was observed when the prime was masked. Interestingly, they also
observed impaired conscious priming but normal subliminal
priming in patients with schizophrenia who exhibited a
hypoactivation of the ACC. These authors concluded that
subliminal conflicts are resolved without ACC involvement. It
seems logical to conclude that if the control structures of the brain
do not play a role in subliminal priming, they are not likely to
participate in conflict adaptation effects. However, in a recent
experiment using fMRI, Blais and Bunge (2010) found that many
of the regions involved in cognitive control in the Stroop task,
including the ACC and DLPFC, were operative at a local item
level.
In the literature on electroencephalography (EEG), ACC
activation has been associated with the N2 component, a negative
deflection in the averaged event-related potential (ERP) wave
with a fronto-central scalp distribution that peaks around 250–
350 ms after stimulus presentation (Ridderinkhof, Ullsperger,
Crone, & Nieuwenhuis, 2004). Consistent with the role of the
ACC in the conflict-monitoring theory, the amplitude of this
component has been found to be more negative in incompatible
trials compared to compatible trials (Clayson & Larson, 2011;
Wendt & Luna-Rodríguez, 2009). Moreover, the P3 component, a
positive deflection in the averaged waveform with a central–
SE I: EXPERIMENTO 1
76
parietal distribution that occurs 350–500 ms after stimulus
presentation, has been associated with response conflict and has
been found to be more positive in incongruent trials compared to
congruent trials (Clayson & Larson, 2011; Frünholz, Godde, Finke,
& Herrmann, 2011). Research conducted by Clayson and Larson
(2011) is of particular interest for the present study. In a flanker
task, these authors studied the sequential effect of the previous
trial on the N2 and P3 components as a function of its
congruency (Gratton, Coles, & Donchin, 1992). This effect is
considered to reflect conflict adaptation and seems to be closely
related to the proportion con- gruency effect (Botvinick et al.,
2001). Clayson and Larson’s results showed that in both
components (i.e., N2 and P3), the difference in amplitude
between an incongruent and a congruent trial was lower when the
previous trial was incongruent than when it was congruent. As for
the ISPCE effect, Shedden, Milliken, Walter, and Monteiro (2012)
analyzed various ERPs in a recent study and failed to find a scalp
signature directly related to the conflict adaptation effect both in a
global–local task and in a standard Stroop task. However, these
authors observed that subjects distinguished the proportion
congruency cat- egory before the congruency of the specific
stimulus as early as 100 ms post-stimulus onset in the global–local
identification task and 150 ms in the Stroop task.
To the extent that unconscious conflict adaptation may
also involve the activation of control structures in the brain, we
expected to find a conflict adaptation effect in a late (i.e., N2 or
P3) event-related component in our experiment. In a recent
UNCONSCIOUS CSPCE IN A STROOP LIKE TASK
77
experiment, Jiang, van Gaal, Bailey, Chen and Zhang (2013)
reported an unconscious block-wise proportion congruency effect
(PCE) in N2 and P3 amplitudes using a meta-contrast masked
priming task. However, to our knowledge, no unconscious ISPCE
or CSPCE has been observed yet. On the other hand, if proportion
congruency is processed earlier than the congruency of the
specific stimulus, as reported by Shedden et al. (2012), we should
expect a similar effect in our experiment since participants are
usually unaware of the proportion of congruency manipulation
even under regular non-masked conditions.
In summary, the goals of our current research were to
explore the Stroop effect under unconscious conditions and to
observe how the proportion of congruency modulates this effect
when it is associated to a particular context. We also intended
to study the time course of neural activation related to processing
of the proportion of congruency, the congruency between the
color word and the color of the target, and the interaction
between item congruency and proportion of congruency that
defines the CSPCE. We focused on analyzing the unconscious
situation by using an extended set of color-word masked trials.
However, we tried to maintain an expectation of the presence of
a color word between masks by calibrating the duration of the
color word individually for each participant until it was unnoticed
and by including a trial block in which the color word could be
easily seen at the end of each session. It has been recently
reported (Martens, Ansorge, & Kiefer, 2011) that subliminal
semantic priming can be modulated by attentional task set. In
SE I: EXPERIMENTO 1
78
this experiment we tried to keep the Stroop task set constant
while the processing of the color word remained unconscious.
2. Material and methods
2.1. Participants
Twenty-nine students of the University of Granada (21
females and 8 males) participated in a two-session experiment.
Sessions were separated by a one-week interval. Participants
reported having normal or corrected-to-normal vision and were
not aware of the purpose of the experiment. Mean age was 23
years (SD = 4.26). Participants received course credit in
exchange for their participation and signed a consent form
approved by the local Ethics Committee.
2.2. Stimuli and apparatus
Three different types of stimuli were presented: three
colored prime words, two different masks and three colored rect-
angles as targets. The color words were GRIS (GREY), ROJO
(RED) or AZUL (BLUE), and the masks could be either @@@@
(visual angle: 4.581º wide by 0.802º high) or #### (visual angle:
6.867º wide by 0.802º high). Each mask was associated with a
different percentage of congruency condition. Words and masks
were presented in 18 Courier New letter font in white on a black
background; words subtended a visual angle 3.437º wide and
0.802º high. Targets were colored rectangles with a visual angle
8.008º wide and 2.864º high. They could appear in gray, red or
blue in the center of the screen. All stimuli were presented on the
UNCONSCIOUS CSPCE IN A STROOP LIKE TASK
79
17-in. color screen of a PC at a viewing distance of about 50 cm.
The PC was connected to a Macintosh com- puter to record
ERPs. The entire task was created and displayed using E-Prime
1.2 Professional software (Schneider, Eschman, & Zuccolotto,
2002).
2.3. Procedure and design
2.3.1. Procedure
The experiment consisted of two sessions separated by a
one-week interval. Both sessions had the same structure. Each
session had two parts: a threshold setting stage followed by an
experimental section. The awareness threshold was deter- mined
for each participant as follows: first, a cross served as a fixation
point and was presented in the center of the screen for 1000 ms.
Next, a color word immediately preceded and followed by the
same mask appeared. The preceding mask was dis- played for
109 ms and the following mask was displayed for 689 ms. At the
beginning, the word was presented for 100 ms, a time interval that
participants could easily perceive. Participants had to report
whether or not they were able to see the word. If the answer was
‘‘yes’’, the presentation time of the word was shortened in 10 ms
steps (equivalent to one refresh rate of the computer screen). If
the answer was ‘‘no’’, the presentation time of the word was kept
constant and the procedure was repeated until the answer
remained ‘‘no’’ for three consecutive trials. The presentation
time obtained was set as the awareness time threshold for the
particular subject. The second part of each session consisted of
SE I: EXPERIMENTO 1
80
four seventy-two masked trial blocks with the following event
sequence: a fixation point appeared for 1000 ms and was
followed by the first mask, which lasted 109 ms. Next, one of
the three color words was presented for the time determined in
the previous part of the experimental session and was followed
by the same mask, which lasted 689 ms. The stimulus-onset
asynchrony (SOA) value between the second presentation of the
mask and the target stimuli was 1000 ms. The target stimuli were
color rectangles in either grey, red or blue ink, which were
displayed until the participant responded by pressing the 1, 2 or 3
num- ber keys of a Qwerty keyboard for gray, red and blue,
respectively. Participants’ response was followed by an intertrial
inter- val that varied randomly between 1000 and 1500 ms (see
Fig. 2.1A).
After the four masked blocks, all subjects performed an
additional block where the word was presented for 100 ms. The
main purpose of this unmasked block was to make subjects
aware that a color word was presented between the two masks,
thus maintaining the expectation of a word between masks as part
of the task set throughout the experiment.1 At the beginning of
the entire procedure we instructed participants to concentrate on
the mask and respond as fast as possible to the color of the
rectangle.
1 Since the main purpose of this experiment was to explore unconscious conflict adaptation, we used an extended number of masked trials but did not use the same number of unmasked trials to avoid discouraging participants. Therefore, data from the two unmasked blocks are not reported.
UNCONSCIOUS CSPCE IN A STROOP LIKE TASK
81
Figure 2.1. (A) Sequence of events in an incongruent trial. Words have
been translated to English. (B) Groups of electrodes selected for P2
(circles), P3 (squares) and N2 (rhombuses) components. Topographies of
the average temporal windows of analysis show the maximum level of
amplitude for each group of electrodes.
2.3.2. Design
Within each block of 72 masked trials two independent variables
were manipulated, namely congruency (congruent vs. incongruent
trials) and percentage of congruency (33% vs. 66%). Each
percentage was associated with a particular mask (#### or
@@@@). In our 33% condition, 12 trials were incongruent and
24 were congruent. On the other hand in the 66% condition, 24
trials were incongruent and 12 were congruent. Thus, the
number of congruent and incongruent trials remained the same
within each block and each mask appeared an equal number of
times but was associated with a different percentage of
congruency.
SE I: EXPERIMENTO 1
82
2.4. EEG recording and data analysis
Participants were seated in front of the computer monitor
in an electrically shielded room. They were instructed to avoid
eye blinking and moving during stimulus presentation and
response. The EEG was recorded using a high-density 128-
channel Geodesic Sensor Net (Tucker, 1993). The head coverage
included sensors lateral to and below both eyes to mon- itor
horizontal and vertical eye movements. The EEG net was
connected to an AC-coupled, high-input impedance amplifier
(200 MX). At the beginning of the recording session, the
impedance of each channel was measured and kept under 50 kX,
as recommended for Electrical Geodesics high-input impedance
amplifiers. Amplified analog voltages (0.1–100 Hz band pass)
were digitized at 250 Hz (12-Bit A/D converter and 0.02 lV
minimum resolvable voltage). All channels were referenced to
the Cz electrode during the recording and were algebraically re-
referenced to the average off-line.2
The continuous EEG was filtered offline using a 30 Hz low-
pass filter. Next, it was segmented in epochs of 200 ms before and
600 ms after target onset. A 200 ms segment previous to the target
presentation was used to calculate the baseline. The epochs were
submitted to software processing in order to identify artifacts.
Segments with eye movements, eye blinks (i.e., electro-oculogram
channel differences greater than 70 lV), more than 20% of bad
2 Following the suggestion made by one reviewer, we also re-referenced all channels to a linked mastoid reference. The outcome of the data analysis using this reference is presented in the results section.
UNCONSCIOUS CSPCE IN A STROOP LIKE TASK
83
channels or incorrect responses were not included in the ERPs.
Data from consistent bad channels were later replaced using a
spherical interpolation algorithm (Perrin, Pernier, Bertrand, &
Echallier, 1989). A minimum criterion of 30 artifact-free trials
per participant and condition was established to maintain an
acceptable signal-to-noise ratio. A final grand average was
obtained for each condition by pooling subjects’ averages in each
experimental condition. Eight group-average ERP waveforms
were constructed according to session (first vs. second),
congruency percentage (33% vs. 66%), and current congruency
(congruent vs. incongruent). Conditions were equated in number
of trials, F < 1.
To facilitate the selection of spatiotemporal windows for
amplitude analyses, the topographic maps provided by the Net
Station Viewer (EGI, 2008) were used as a guide. Topographic
map views display samples as representations of the scalp
projected onto disk-shaped maps (also referred to as scalp maps)
where amplitudes are represented by colors. In such views, the
amplitudes between sensors are interpolated, which allows the
entire surface of the head to be depicted, thus providing a view of
the voltage distribution (i.e., topographies) over the scalp for
each experimental condition as a function of time. As shown in
Fig. 2.1B, visual analysis of the scalp maps revealed two main
focuses of activity in the ERPs: a posterior focus, which included
seven parieto-occipital electrodes (shown as circles) and remained
for 205–275 ms after target presentation, and a second more
central focus, which included thirteen parietal electrodes (shown
SE I: EXPERIMENTO 1
84
as squares) and remained for 360–400 ms after the presentation
of the target. These two topographies corresponded to the P2 and
P3 components. Voltage analysis was performed on a 40 ms
window centered at the P2 peak (240 ms) of the grand average
waveform for the posterior location and on a 20 ms window
centered at the P3 peak (380 ms) for the central location. Using a
mastoid reference, the scalp maps revealed an additional focus of
activity in the ERPs: a frontal focus, which included twelve
electrodes (shown as rhombuses) centered on the midline and
remained for 260–300 ms after target presentation. This
topography corresponded to a N2 component. Voltage analysis
was performed on a 40 ms window centered at the N2 peak (280
ms) of the grand average waveform.3 Trials that did not meet the
criteria regarding the amount of artifacts and bad channels were
eliminated for each participant. Mean amplitude voltages averaged
over the selected channels and time windows were submitted to
repeated-measures ANOVAs with Session, Congruency Percentage
and Congruency as factors. The same ANOVA was used with
subjects’ accuracy and RTs.
3 After analyzing this twelve-channel focus using a mastoid reference, we also analyzed the N2 component of the twelve channels using an average reference. Results are presented in the corresponding section
UNCONSCIOUS CSPCE IN A STROOP LIKE TASK
85
3. Results
3.1. Behavioral results
3.1.1. Threshold setting
The average presentation time of the masked words at
threshold value was 15.97 ms (SD = ± 5.53 ms) in the first session
and 16.66 ms (SD = ± 8.03 ms) in the second session. A
comparison between these two threshold values showed no
significant differences between sessions, t = .586, p = .562.
3.1.2. RT analysis
Overall accuracy was very high (99%) and no difference in
accuracy was observed between experimental conditions. Only
correct responses were considered in the RT and ERPs analyses.
RTs shorter than 300 ms or longer than 1000 ms were excluded as
outliers (3%).
No significant main effects were found in the analysis of
RTs; however, the interaction between Congruency Percentage and
Congruency was significant, F(1, 27) = 5.014, p < .05 (see Fig.
2A). Planned comparisons showed that, when the propor- tion of
incongruent trials was low (33%), responses to incongruent trials
(mean = 525 ms, SD = ± 43.5 ms) were slower than responses to
congruent trials (mean = 515 ms, SD = ± 49 ms), F(1, 27) = 4.8, p
< .05. By contrast, no difference in RTs was observed between
congruent and incongruent trials (mean = 522.25 ms, SD = ± 41.75
ms) when the proportion of incongruent trials was high (66%),
SE I: EXPERIMENTO 1
86
F(1, 27) = .02, p > .1. Thus, the expected conflict adaptation effect
was found in RT measures.
As some authors have argued that the proportion
congruency effect is related to the conflict adaptation sequential
effect (e.g., Botvinick et al., 2001), we further analyzed RT data
including the congruency or incongruency of the previous trial as a
new independent variable. No significant effects were found.
3.2. Electrophysiological data
3.2.1. Target analysis
As mentioned above, inspection of the scalp topographies
revealed a P2 positive polarity lasting from 205 to 275 ms after
target presentation over a set of parieto-occipital electrodes. As
shown in Fig. 2B, waveforms for the percentage of congruency
conditions showed a difference in amplitude that peaked around
240 ms. A window of 80 ms around this peak was analyzed.
UNCONSCIOUS CSPCE IN A STROOP LIKE TASK
87
Figure 2.2. (A) Interaction of Incongruency percentage x Congruency
effect on RT. (B) Waves and topographies of the main effect of
Incongruency percentage on the P2 amplitude. (C) Waves and
topographies corresponding to the interaction of Incongruency
percentage x Congruency on the P3 amplitude. The 33% effect on
Congruency appears on the left (33% congruent in gray and 33%
incongruent in black) while the 66% effect on Congruency appears on the
right (66% congruent in gray and 66% incongruent in black). In the center
we present the amplitude table for this interaction.
SE I: EXPERIMENTO 1
88
The amplitude in the 66% incongruent trial condition
(mean 66% = 1.459, SD = ± 2.046) was higher than the amplitude
in the 33% incongruent trial condition (mean 33% = 1.239, SD
= ± 2.042). This difference was found to be significant, F(1, 27)
= 7.206, p < .05. No effect of the congruency variable was found
in this time window. This main effect of the percent- age of
congruency was not significant when a linked mastoid reference
was used.
A second, later positive polarity (P3) was identified
between 360 and 400 ms at a more central position encompassing
thirteen electrodes. An amplitude analysis of this window showed
no significant main effect but revealed a significant inter- action
between congruency percentage and congruency, F(1, 27) = 9.622,
p < .01. As shown in Fig. 2C, the amplitude was higher in
incongruent trials (mean 33% I = 4.209, SD = ± 2.168) than in
congruent trials (mean 33% C = 3.825, SD = ± 2.118) when the
percentage of incongruent trials was low (33%), and this difference
was significant, F(1, 27) = 6.508, p < .05. By contrast, this trend
was reversed when the percentage of incongruent trials was high
(66%); in this case, the amplitude was higher in incongruent trials
(mean 66% I = 3.897, SD = ± 2.003) than in congruent trials (mean
66% C = 4.052, SD = ± 2.189), although this difference did not
reach significance, F > .1.
UNCONSCIOUS CSPCE IN A STROOP LIKE TASK
89
Figure 2.3. Waves and topographies corresponding to the interaction of
Incongruency percentage x Congruency on the N2 amplitude. The 33%
effect on Congruency appears on the left (33% congruent in gray and
33% incongruent in black) while the 66% effect on Congruency appears
on the right (66% congruent in gray and 66% incongruent in black). In
the center we present the amplitude table for this interaction.
This interaction was also significant when a mastoid
reference was used, F(1, 27) = 4.403, p < .045. It was mainly
produced by a significant decrease in amplitude in incongruent
items from the 33% condition (mean = 4.72, SD = ±.438) to the
66% condition (mean = 4.31, SD = ±.382), F(1, 27) = 4.403, p < .05.
Thus, we also found a conflict adaptation effect in this amplitude
measure.
A significant frontal effect on N2 (see Fig. 2.3) was found
when an average reference was used as revealed by the interaction
SE I: EXPERIMENTO 1
90
between percentage of congruency and congruency F(1, 27) =
4.227, p < .05. Only the comparison between the 33% condition
(mean = -2.50, SD = ±.64) and the 66% condition (mean = -
3.05, SD = ± 1.11) for congruent items was significant F(1,
27) = 8.194, p < .01. Using a mastoid reference, the analysis
revealed a marginally significant interaction between per-
centage of congruency and congruency, F(1, 27) = 3.170, p <
.086. Only the comparison between the 33% condition (mean
= -2.86, SD = ±.81) and the 66% condition (mean = -3.67, SD
= ± 1.66) for congruent items was significant, F(1, 27) = 5.13,
p < .032.
3.2.2. Mask analysis
Throughout the experiment, each mask was systematically
associated with a particular percentage of congruency in order to
determine the extent to which the brain was registering this
association. Therefore, we further analyzed the ERPs associated
with the presentation of the mask following the color word. It
should be kept in mind that, given the close proximity between the
color word and the mask, the processing of such stimuli was likely
to overlap, making it impossible to attribute the ERP waveform
only to the mask. Consequently, our intention with this analysis
was to determine whether or not the brain had differentially
registered the association between each mask (or word-mask
compound) and its corresponding pro- portion of congruency.4
4 From now on we shall refer only to the mask.
UNCONSCIOUS CSPCE IN A STROOP LIKE TASK
91
The scalp maps indicated one main focus of posterior
activity related to the mask that included the same group of parieto-
occipital electrodes analyzed for the target. Using a baseline of
200 ms before the second mask presentation, we analyzed two
temporal windows: a window between 210 and 250 ms, peaking
around 230 ms, and a window between 270 and 345 ms, peaking
around 300 ms. In the first temporal window, a main effect of
congruency percentage was found to be significant, F(1,) = 8.38,
p < .01, showing a higher positive deflection in the mostly
incongruent condition (mean = 1.92, SD = ± 2.055) compared to
the mostly congruent condition (mean = 1.67, SD = ± 2.085). We
also found an interaction close to significance between session and
congruency percentage, F(1, 27) = 3.93, p = .057. Planned
comparisons showed a significant difference in amplitude between
both levels of congruency percentage only in the second session,
with higher results in the mostly incongruent condition (mean =
2.160, SD = ± 2.14) compared to the mostly congruent condition
(mean = 1.765, SD = ± 2.145), F(1, 27) = 12.507, p < .01.
4. Discussion
The main purpose of the present research was to explore
whether or not the proportion of congruency modulated the
Stroop effect when the color words were masked. Both RT and ERP
measures provided an affirmative answer to this question. RTs in
congruent and incongruent trials significantly changed depending
on the percentage of incongruent trials. When this percentage was
low, congruent trials were responded to faster than incongruent
SE I: EXPERIMENTO 1
92
trials; by contrast, when the percentage of incongruent trials was
high, no significant differences in RTs were found between these
two conditions. In the ERPs, the amplitude of both the N2 and the
P3 components showed a significant interaction between the
proportion of congruency and the congruency between the color
word and the color of the target.
We applied a subjective threshold of awareness
determined by presentation time. Below this threshold, subjects
were unable to perceive the color word. In fact, mean
presentation time was 15.97 ms in the first session of the
experiment and 16.66 ms in the second session, with no
significant differences in threshold values between both. These
presentation times were similar or even lower than those
generally used in other masking experiments (Daza et al., 2002;
Greenwald, Klinger, & Liu, 1989; Klapp, 2007, Ortells, Marí-
Beffa, & Plaza-Ayllón, 2002). As mentioned above, the CSPCE has
been found to be independent of subjects’ awareness of the
proportion of congruency even in situations where no masking
procedure was used (Crump et al., 2008). In the present study,
subjects were unaware of the congruency relation between the
masked color word and the color of the patch they had to name; in
addition, the percentage of congruent and incongruent trials was
the same within each block of trials. Under these conditions,
awareness of the proportion of congruency associated with each
mask seems very unlikely. After the experiment, when subjects
were asked if they had been able to process any color words, the
general answer was ‘‘no’’ and none were able to establish any
UNCONSCIOUS CSPCE IN A STROOP LIKE TASK
93
relationship between the masks and the proportion of con-
gruency. Moreover, all subjects reported the same difficulty across
blocks regardless of the mask type. Based on these results, we can
conclude that the context-specific proportion congruency effect
(CSPCE) was obtained at least under conditions in which
participants were subjectively unaware of the congruency relation
between a color word and a color patch in a modified Stroop task.
Whether or not this effect can also be obtained under objective
thresholds of consciousness requires further study. Nevertheless,
the fact that the CSPCE was obtained under subjective threshold
conditions should not be underestimated. The subjective
threshold is likely to be the main point to determine awareness
and provides relevant information about cognitive processing
outside the realm of consciousness, as shown by research on
various neuropsychological deficits, such as blindsight
(Weiskrantz, 2009) or spatial neglect (see Driver & Vuilleumier,
2001).
The analysis of the event-related potentials associated with
the target provided additional information to the RT results. First,
our main finding was related to a P3 (or P300) component that
peaked around 380 ms and showed the expected inter- action
between percentage of congruency and congruency that
characterizes the CSPCE. The amplitude of incongruent trials was
significantly higher than that of congruent trials in the mostly
congruent condition; by contrast, this difference in amplitude
disappeared in the mostly incongruent condition. This interaction
between percentage of congruency and congruency was obtained
SE I: EXPERIMENTO 1
94
using an average and a mastoid reference. The P3 component has
been associated with response conflict and reported to be more
positive in incongruent compared to congruent trials (Clayson &
Larson, 2011; Frünholz et al., 2011). Although this component has
also been associated with conscious processing (Dehaene &
Changeux, 2011), our results show that it may be affected by the
unconscious processing of both the proportion of congruency and
the congruency between the color word and the color of the
target. An influential though highly controversial interpretation of
the P300 component (Donchin & Coles, 1988) is that this
component is a manifestation of activity occurring whenever an
individual’s environ- mental context must be updated. Our results
fit well with this view and provide the additional information that
this ‘‘context updating’’ need not be conscious. In our study, each
mask set the context for a different proportion of congruency to
influ- ence the congruency effect between the color word and the
color of the target. It has been argued that the CSPCE requires
conscious representation of conflicting information, namely the
prime, target and context (Kunde et al., 2012). However, our
results indicate that a conscious representation of the prime is not a
necessary condition for the CSPCE to take place.
Second, a significant interaction between percentage of
congruency and congruency in a frontal N2 when an average ref-
erence was used and the mastoid reference analysis yielded a
marginally significant interaction. Therefore, this component also
showed evidence that the proportion of congruency associated to
UNCONSCIOUS CSPCE IN A STROOP LIKE TASK
95
the corresponding mask modulated the congruency between the
color-word and the target.
Third, at least in the average-reference analysis, a relatively
early P2 component around 240 ms showed a difference in
amplitude as a function of the proportion of incongruent trials.
Specifically, the amplitude of mostly incongruent trials was higher
than that of mostly congruent trials. The proportion of congruency
seems to have already been processed at this point within the
wavelength and thus to set the stage for a later appearance of the
CSPCE.
In summary, the present research revealed that the CSPCE
was clearly present in the P3 and N2 components. Both com-
ponents have been associated with the operation of control
structures such as the ACC and DLPFC (Frünholz et al., 2011;
Ridderinkhof et al., 2004). It has been demonstrated that these
structures play a role in conflict adaptation at a local item level
(Blais & Bunge, 2010). Our results add further support to the
findings of Blais and Bunge. They suggest that, at least under the
conditions of the present research, these control structures can
operate even when participants are unaware of the relationship
between the color word and the color of the target and the
relationship between the masks and the proportion of congruency.
Finally, the analysis of the waveform associated with the
mask yielded a significant effect of proportion of congruency on
the posterior P2. Specifically, the response to the mask associated
with mostly incongruent trials had greater amplitude than the
SE I: EXPERIMENTO 1
96
response to the mask associated with the mostly congruent trials.
Interestingly, this P2 difference in amplitude increased across
sessions, reaching significance in the second session. Thus, it
seems that the building of an association between a given mask
and a particular percentage of congruency takes place across trials
and implies some kind of implicit or unconscious learning. In fact,
Wong, Bernat, Snodgrass, and Shevrin (2004) have reported effects
of associative learning on the P2 component.
Taken together, our results show that, even though
subjects were unconscious of the congruency relation between the
masked color word and the color of the target and of the
proportion of congruency associated with each mask, the brain
responded sensitively to both variables and to their interaction,
giving rise to the CSPCE. The presentation of the mask probably
set the context that modulated the congruency effect produced by
the relationship between the color word and the color of the
target. The interaction between percentage of congruency and
experimental session found in the P2 of the mask waveform
suggests that the building of an association between a given mask
and a particular percentage of congruency takes place across trials
and implies some kind of implicit or unconscious learning.
However, the CSPCE seems to be a within-trial process where the
context activated by the mask modulates the congruency response.
This account of the CSPCE is consistent with the explanation of the
item-specific proportion congruency effect (ISPCE) provided by
Shedden et al. (2012). According to these authors, control
UNCONSCIOUS CSPCE IN A STROOP LIKE TASK
97
processes can be implemented rapidly (i.e., online) based on a
previously experienced conflict in a particular context.
It should be kept in mind that throughout the present
experiment great care was taken to ensure that the expectation of a
color word between the two masks remained activated. In other
words, subjects were generally conscious that a color word was
being presented between the masks, although they were not
conscious of which exact color word was presented in each trial.
Attentional modulation of subliminal semantic processes by task
set has recently been reported (Martens et al., 2011; see also
Kiefer, Adams, & Zovko, 2012). The general expectation of the
presence of a color word between masks may have contributed to
sensitizing subjects about the semantic representation of the colors
used in our task even though they were not aware of the
particular color word presented in each trial. This explanation
may account for the congruency effect between the color word
and the color of the target but cannot explain the relationship
between the masks and the proportion of congruency. In this case,
the relationship could not have been conscious; instead, it appears
to have been built through practice and remained unconscious all
along the experiment. However, such relationship exhibited a
clear effect both on RT and ERP measures. In addition, the
CSPCE observed in our experiment did not seem to be
attributable to sequential effects, as no significant effects were
found when we included the congruency of the previous trial as a
variable in the analysis of our RT data.
SE I: EXPERIMENTO 1
98
Overall, our results lend support to the currently widely
held view that the classical division between automatic and con-
trolled processes (Posner & Snyder, 1975; Schneider & Shiffrin,
1977; Shiffring & Schneider, 1977) is in need of a thorough
revision (Bugg & Hutchison, 2013; Crump et al., 2008; Shedden et
al., 2012). According to this division, controlled and automatic are
conceptualized as categories belonging to a general and
disjunctive classification of processes that makes it contradictory
even to talk about automatic control. Each category is defined by a
set of characteristics. For example, automatic processes are
considered to be fast, stimulus-driven, resource-free and to occur
in the absence of awareness of the stimuli that elicit them. By
contrast, controlled processes are thought to be slow, subject to
volition, resource demanding and open to conscious inspection.
This way of classifying cognitive processes is incompatible with
our results, which show control adaptation in a situation where
subjects are unaware of the relationship between a mask and a
proportion of congruency and of the congruency relation between
a color word and the color of the target. By contrast, an influential
model of control processes (Norman & Shallice, 1986; see also
Shallice & Cooper, 2011) allows for a more flexible relationship
between two main components: (1) a contention scheduling
mechanism composed of a set of automatic schemas resulting from
experience and learning that maintain cooperative or competitive
relations among themselves, with the most active schema taking
con- trol of a particular situation; and (2) a supervisory attentional
system (SAS) responsible for conscious control that comes into
UNCONSCIOUS CSPCE IN A STROOP LIKE TASK
99
play in situations where routine schemas cannot be relied upon.
According to this way of thinking, contention scheduling and the
SAS are not ways of classifying cognitive processes. Instead, they
are conceptualized as two different systems that may act
cooperatively to achieve cognitive control. Under this view, it
makes perfect sense to talk about automatic control when a
particular schema exerts control as a result of learning. In our
study, as a consequence of the association between a mask and a
particular proportion of congruency, an unconscious control
schema was generated; this schema was able to act in the presence
of that particular mask, thus modulating the Stroop effect, which in
turn took place without awareness of the color word. In addition,
as in the Norman and Shallice model, we found that the
relationship between contention scheduling and the SAS is
cooperative rather than disjunctive, allowing the flexible
attentional top-down influence of the conscious task set, that
remained constant in our experiment, on unconscious processing
as recently elaborated by Kiefer et al. (2012).
5. Conclusions
Our results provide evidence of the unconscious character
of cognitive control when a masked Stroop-like paradigm is used.
The early P2 component showed amplitude differences as a
function of percentage of congruency, displaying larger amplitude
for mostly incongruent than for mostly congruent trials. This
seems to be evidence of a type of implicit learning across trials
SE I: EXPERIMENTO 1
100
given that, when masks were analyzed, the P2 tended to be higher
in the second session compared to the first one.
The interaction between the percentage of congruency and
the congruency between the color word and the color of the target
observed in the N2 and P3 component and RTs provided evidence
of the expected CSPCE even when subjects were unaware of two
main relationships in the experimental task: the relationship
between the color word and the color of the target, and the
relationship between the masks and the proportion of congruency.
Acknowledgments
We thank María Ruz and our reviewers for their careful
reading of our manuscript and their many insightful comments and
suggestions.
This research was supported by the Spannish Ministerio de
Educación y Cultura with a Grant (FPU-AP2007-00313) to the first
author and a research Grant P09-HUM-5422 from the Andalusian
Autonomous Government to Pío Tudela..
Capítulo 3
SERIE EXPERIMENTAL II:
Comparing conflict adaptation in cognitive
and emotional conflict
1. Introduction
As well as in real life activities, when performing an
experimental task in which there is competition between the task-
relevant and task-irrelevant information, humans need to implement
control mechanisms (see Norman & Shallice, 1986; Fan & Posner,
2004). How these control mechanisms are implemented and how
our brain identifies this kind of competing information and puts into
action strategies to avoid irrelevant information is a question that
has been studied for over the years. For this purpose, the stimulus-
response compatibility tasks (Kornblum, Hasbroucq, & Osman,
1990) have been frequently used. A well known variant of this type
of tasks is the Stroop task (Stroop, 1935; Macleod, 1991) in which
participants have to name the ink color of color words. There are
two conditions: congruent and incongruent. In the congruent
condition, the ink color matches the color name of the word (e.g.,
RED in red ink). In the incongruent condition, by contrast, the ink
color is different from the color name of the word (e.g., BLUE in
red ink). To perform the task in the incongruent condition,
participants must avoid the automatic process of reading the word
and name the color of the ink in which the word is printed by using
control strategies. In the congruent condition, the ink color matches
the color name of the word, so avoidance of the automatic reading
process is not necessary. Consequently, reaction times (RTs) are
longer in the incongruent condition than in the congruent one. This
additional time is known as Stroop interference or Stroop effect.
SE II: EXPERIMENTO 1
104
On the other hand, when a sequence of trials within a block
is considered, a conflict adaptation effect appears (Gratton, Coles &
Donchin, 1992) as a reduction in interference in a current
incongruent trial when it is preceded by another incongruent trial.
This effect has been found across different conflict tasks, including
Stroop (e.g., Egner & Hirsch, 2005), Simon (Hommel, Proctor &
Vu, 2004) and flanker tasks (e.g., Nieuwenhuis, Stins, Posthuma,
Polderman, Boomsma & De Geus, 2006); it seems to be domain-
specific, as it does not transfer across different tasks that are
performed sequentially (Egner, 2008; Funes, Lupiáñez, &
Humphreys, 2010).
In computational and neural terms, Botvinick, Braver,
Barch, Carter & Cohen (2001) advanced an explanation of both, the
Stroop and the conflict adaptation effects, in what they termed the
conflict-monitoring hypothesis. According to these authors, the
Stroop effect is just a particular case of conflict resolved by the
interplay of a set of neural structures. The dorsal anterior cingulate
cortex (dACC) acts as a detection mechanism that responds to the
occurrence of a conflict situation (i.e., an incongruent trial in a
Stroop task). In addition, this conflict signal triggers strategic
adjustments in the dorsolateral prefrontal cortex (DLPFC) that
amplifies cortical responses to reduce conflict. On the other hand,
the conflict adaptation effect is considered to result from the
activation of the DLPFC that serve to prevent conflict in subsequent
performance. Thus, according to the conflict monitoring hypothesis,
when an incongruent trial is presented dACC detects the conflict
situation triggering signals to DLPFC that implements strategies to
COMPARING CONFLICT ADAPTATION IN CONFLICT
105
resolve this situation. If a subsequent incongruent trial is presented,
the DLPFC is now prepared for the resolution of the conflict, which
translates in faster RT in the current incongruent trial. Experiments
using functional magnetic resonance (fMRI) have lent support to
this conflict-monitoring hypothesis (Botvinick, Cohen, & Carter,
2004; Kerns, Cohen, McDonald III, Cho, Stenger & Carter, 2004;
McDonald, 2000).
However conflict not only exists in the cognitive but also in
the emotional domain. Many tasks have been adapted in order to
study the emotional conflict resolution processes (Bishop, Duncan,
Brett, & Lawrence, 2004; Bush, Luu, & Posner, 2000; Compton,
2003; Shin, Whalen, Pitman, Bush, Macklin, Lasko et al., 2001;
Whalen, Bush, Mcnally, Wilhem, McInerney, Jenike & Rauch,
1998). For example, in the emotional counting Stroop task (Whalen
et al., 1998), participants have to report the number of words that
appear on a screen, regardless of the word meaning. Control trials
contain emotionally neutral words (e.g., ‘cabinet’ written three
times), while interference trials contain emotional words (e.g.,
‘murder’ written three times). As a general finding, interference
(emotional) words tend to show greater interference effect (longer
RT) when compared with control (non-emotional) words. Similar
results have been found in emotional Stroop tasks where subjects
have to name the ink color of words the meaning of which may be
emotional or non-emotional. As in the previous task, emotional
words produce longer RT than non-emotional words (e.g. Watts,
McKenna, Sharrock & Trezise, 1986).
SE II: EXPERIMENTO 1
106
A common characteristic of these tasks is that emotional
information acts as distracter rather than target, because target
information is always cognitive. Thus they may be suited to
provide information about the interference produced on the control
of a cognitive activity (either counting or color naming) by
emotional information, but they do not provide direct information
on emotional control. To achieve this information, emotional
stimuli should be included as targets. Recently, attempts have been
made to study both, cognitive and emotional, control mechanisms
within the same task in order to equate the cognitive and affective
type of conflict. The face-word Stroop task (Egner, Etkin, Gale, &
Hirsch, 2008; Etkin, Egner, Peraza, Kandel, & Hirsch, 2006) may
be considered a good example. In the emotional version of this task,
subjects have to name the emotion displayed by either a male or
female face while an emotional word is presented across the face
such that the meaning of the word and the emotion of the face may
be either congruent or incongruent. In the cognitive version of the
task, subjects have to respond to the gender (male or female) of the
faces disregarding the meaning of the word (male or female),
printed across the face, that may be congruent or incongruent with
the face gender. Results show that in both versions of the task the
incongruent condition creates conflict (emotional or cognitive),
reflected in slower RT in incongruent relative to congruent trials.
Other researchers (Alguacil, Tudela, & Ruz, 2013; Ochsner,
Hughes, Robertson, Cooper, & Gabrieli, 2008) have employed a
variant of the Eriksen flanker task (Eriksen & Eriksen, 1974) in
which a word was presented as target at the center of the screen and
COMPARING CONFLICT ADAPTATION IN CONFLICT
107
flanked by another two words above and below. In the emotional
version, participants had to respond to the valence of the target
(positive or negative); in the cognitive version they had to indicate
the target’s semantic category. In either version, the flanker words
could be congruent or incongruent with the target. As in the Stroop
task also in the flanker task the incongruent trials produced both,
cognitive and emotional conflict (slower RT).
Regarding neural mechanisms underlying cognitive and
affective conflict, an influential review of experiments, using
imaging techniques, on cognitive and emotional versions of the
Stroop task (Bush, Luu, & Posner, 2000) suggested a segregation of
functions within the anterior cingulated cortex (ACC); the dorsal
part would be mainly involved in cognitive control while the ventral
ACC (vACC) would be engaged when emotional control is needed.
This hypothesis was based in experimental evidence where
cognitive and emotional control was studied in experiments in
which, as pointed out above, emotional information acted as
distracter for cognitive responses. More recently, the neural
substrate of cognitive and affective conflict has been jointly
explored within the same experiment using fMRI (Egner, et al.,
2008; Ochsner, et al., 2008). Their results partially modified Bush et
al.’ hypothesis so that emotional conflict seems to recruit vACC and
rostral medial prefrontal cortex (RMPC) while cognitive conflict
appears under control of the DLPFC.
Although fMRI studies have provided important information
about brain mechanisms involved in conflict resolution, the timing
SE II: EXPERIMENTO 1
108
of the different processes involved cannot be studied with precision.
Event-related potential (ERP) recordings, on the other hand, can
provide critical temporal information for precise analysis of the
timing of different types of conflict. To date, at least one
experiment have studied the ERPs associated to cognitive and
emotional conflict jointly in the same experiment and using the
same kind of material and task settings across conditions (Alguacil,
et al., 2013). In this experiment participants performed a flanker
task in which they had to focus on a central target and indicate its
semantic category (cognitive version) or its valence (affective
version). Targets were flanked by congruent or incongruent words
in both versions. Results showed differences in the wave amplitude
as a function of the cognitive or emotional nature of the task in the
late ERP components but none in the early components. They found
conflict adaptation effects in N170 component with larger
amplitudes for incongruent trials preceded by a previous
incongruent trial. This pattern was the same for both, emotional and
non-emotional task, showing that conflict had been processed at a
very early stage of information processing. They also found that the
N2 and the first part of the P3 were exclusively modulated by
cognitive conflict, whereas the last section of the P3 deflection was
only involved in affective conflict processing. The authors argued
in favor of the existence of early common mechanisms, equivalent
for cognitive and affective materials, and later task-specific conflict
processing mechanisms.
The N170 is an early potential that reflects perceptual
processing in the visual cortices. It has been associated with
COMPARING CONFLICT ADAPTATION IN CONFLICT
109
perceptual discrimination and object categorization processes
(Hillyard, 2009; Luck et al., 2000) and can be modulated by
attention (see Aranda, Madrid, Tudela, & Ruz, 2010; Ruz & Nobre,
2008a). ERP studies using both emotional and non-emotional
conflict tasks have shown larger N170 amplitudes for incongruent
than for congruent trials in both types of conflict (Zhu, Zhang, Wu,
Luo & Luo, 2010). This experiment used a version of the face-
categorization task similar to that introduced by Egner et al. (2008).
When participants were asked to identify the emotional expression
of the face, the incongruent condition evoked a more negative N170
component than in the congruent condition. In contrast, when
subjects were asked to identify the emotional meaning of the word,
the pattern of results reversed. These results provide evidence that
the N170 may be modulated by emotional conflict
The N2 component, that is strongly associated with
cognitive control, is a negative deflection that takes place around
200 ms after stimulus onset with a fronto-central topography. The
amplitude of this component is usually larger for incongruent than
for congruent trials (Koop, Rist, & Mattler, 1996; Van Veen &
Carter, 2002). This result seems to reflect the implementation of
control mechanisms and has been replicated several times (Folstein
& Van Petten, 2008; Van Veen & Carter, 2002). As mentioned
before, the study of emotional conflict has been limited to the
influence of irrelevant emotional material on the resolution of
cognitive conflict. In this type of situation, the presence of
irrelevant emotional material has also generated an N2 of larger
amplitude compared to a control condition (Kanske & Kotz, 2010,
SE II: EXPERIMENTO 1
110
2011). However in the experiment by Alguacil et al. (2013) where
the valence of the stimuli was relevant to performance, no effect on
the N2 component was found in the emotional task though both, the
previous and current congruency variables, significantly modulated
this component in the cognitive task.
As to the P3 component, a positive deflection in the
averaged waveform with a central-parietal distribution occurring
350-500 ms after stimulus presentation, it has been associated with
stimulus selection and categorization and with the implementation
of the required response (Luck, 2005). It has been argued (Polich,
2007) that in conflict situations the P3 may be generated by
inhibitory processes involved in avoiding irrelevant and competing
information , as well as in focusing attentional resources on the
relevant elements of the task. As a matter of fact, Clayson &
Larson (2011) found a more positive P3 for incongruent than
congruent trials (see also Frünholz, Godde, Finke & Herrmann,
2011). Clayson & Larson’s (2011) research is also relevant to the
present investigation because they studied the conflict adaptation
effect in a flanker task and found that in both, the N2 and P3
components, the amplitude of an incongruent trial was reduced
when the previous trial was incongruent compared to congruent.
As the electrophysiological evidence about the relationship
between emotional and cognitive conflict is scant, in the present
experiment we tried to compare, jointly in the same experiment and
using the same kind of material and task settings across conditions,
the ERPs associated with these two types of conflict, as well as with
COMPARING CONFLICT ADAPTATION IN CONFLICT
111
the conflict adaptation effects associated with both types of
conflicting situations. For this purpose we used a face-
categorization task (Egner y Hirsch, 2005b; Gale; and Hirsch,
2007), in which we varied the source of conflict between non-
emotional and emotional stimulus representations, while keeping
task-relevant stimulus characteristics constant. The task-relevant
information for both tasks was a set of male and female faces
expressing either happiness or fear (Ekman & Friesen, 1976). For
the non-emotional task, subjects had to respond to the gender of the
face while ignoring gender labels appearing across the faces. For
the emotional task subjects had now to respond to the emotion
expressed by the face ignoring a congruent or incongruent
emotional label across the faces. During the entire experiment we
recorded continuously the EEG of the participants.
2. Material and methods
2.1. Participants:
Twenty four participants (16 female and 8 male) students of
the University of Granada participated in a one-session experiment.
They reported to have normal o corrected-to-normal vision, and
they were not aware of the purpose of the experiment. The mean
age was 25 years (SD= 6.30). All of them gave their verbal consent
to participate in the experiment and signed a consent form approved
by the local Ethics Committee. Credits were given for participation.
SE II: EXPERIMENTO 1
112
2.2 Stimuli and apparatus
The stimulus set consisted of 5 male and 5 female faces with
an expression of fear or happiness (11.42º high and 10.28º wide of
visual angle). In the non-emotional task faces were presented with
either the word HOMBRE (man) (0.573º high and 4.581º wide of
visual angle) or MUJER (woman) (0.573º high and 3.437º wide of
visual angle) printed in red ink, producing gender-congruent and
incongruent stimuli (see Fig 1. A.). In the emotional version the
same face stimuli were presented with either the word ALEGRIA
(happiness) (0.573º high and 4.581º wide of visual angle) or
MIEDO (fear) (0.573º high and 3.437º wide of visual angle). All
stimuli were presented on a 17 inch color screen of a PC computer
at a viewing distance of about 50 centimeters. The PC was
connected to a Macintosh computer for the registration of the ERPs.
The entire task was created and displayed using the E-Prime 1.2
Professional software (Schneider, Eschman & Zuccolotto, 2002)
2.3. Procedure and design
2.3.1. Procedure
As mentioned before the task had two versions being the
structure of the task the same for each version. First a fixation point
appeared at the center of the screen with a duration varying
randomly between 500 and 1000 ms. Then, the target stimulus was
presented during 1000 ms or until the participant gave a response.
Finally we displayed an inter-trial interval (ITI) varying between
1000 and 1500 ms (see Fig. 3.1).
COMPARING CONFLICT ADAPTATION IN CONFLICT
113
Figure 3.1. Sequence of a non-emotional congruent and emotional incongruent
trial.
Participants used a computer Qwerty keyboard to answer.
They were asked to respond to the gender of the face (hombre or
mujer) in the cognitive version and to the emotion of the face
(alegría or miedo) in the emotional version of the task. In every trial
the face was displayed in the center of the screen with an
overlapping word placed over the nose of the face. The meaning of
the word could be congruent or incongruent either with the gender
of the face (non-emotional version) or with the emotion displayed
by the face (emotional version). Participants were encouraged to
focus on the face and disregard the words as much as possible. They
were repeatedly told to respond based upon the face and to ignore
the word.
For each version of the task we conducted 10 practice trials
and four blocks of 160 trials each. During the practice trials
participants received feedback of their performance. Answer keys,
SE II: EXPERIMENTO 1
114
blocks and the starting version of the task were counterbalance
across participants.
2.3.2. Design
Within each 160 trials block of each version of the task, two
independent variables were manipulated: current trial congruency
(congruent vs incongruent) and previous trial congruency
(congruent vs incongruent) A 2 (previous trial congruency) x 2
(current trial congruency) repeated measures ANOVA was
conducted. Task version was manipulated between blocks and the
order of presentation was counterbalanced per participants.
2.4. EEG recording and analysis:
Participants were seated in front of the computer monitor in
an electrically shielded room. They were instructed to avoid eye
blinking and moving during stimulus presentation and response.
The EEG was recorded using a high-density 128-channel Geodesic
Sensor Net (Tucker et al., 1994). The head coverage included
sensors lateral to and below both eyes to monitor horizontal and
vertical eye movements. The EEG net was connected to an AC-
coupled, high-input impedance amplifier (200 M). At the
beginning of the recording session, the impedance of each channel
was measured and kept under 50 kΩ, as recommended for Electrical
Geodesics high-input impedance amplifiers. Amplified analog
voltages (0.1–100 Hz band pass) were digitized at 250 Hz (12-Bit
A/D converter and 0.02 V minimum resolvable voltage). All
COMPARING CONFLICT ADAPTATION IN CONFLICT
115
channels were referenced to the Cz electrode during the recording
and were algebraically re-referenced to the average off-line.
The continuous EEG was filtered offline using a 30 Hz low-
pass filter. Next, it was segmented in epochs of 200-700 ms relative
to the target onset. A 200 ms segment previous to the target
presentation was used to calculate the baseline. The epochs were
submitted to software processing in order to identify artifacts.
Segments with eye movements, eye blinks (i.e., electro-oculogram
channel differences greater than 70 V), more than 20% of bad
channels or incorrect responses were not included in the ERPs. Data
from consistent bad channels were later replaced using a spherical
interpolation algorithm (Perrin, Pernier, Bertrand, & Echallier,
1989). A minimum criterion of 30 artifact-free trials per participant
and condition was established to maintain an acceptable signal-to-
noise ratio. Trials that did not meet the criteria regarding the
amount of artifacts and bad channels were eliminated for each
participant. A final grand average was obtained for each condition
by pooling subjects’ averages in each experimental condition. Eight
group-average ERP waveforms were constructed according to task
(emotional vs. non-emotional), previous trial congruency
(congruent vs. incongruent) and current congruency (congruent vs.
incongruent). Conditions were equated in number of trials, F 1.
To facilitate the selection of spatiotemporal windows for
amplitude analyses, the topographic maps provided by the Net
Station Viewer (EGI, 2008) were used as a guide. Topographic map
views display samples as representations of the scalp projected onto
SE II: EXPERIMENTO 1
116
disk-shaped maps (also referred to as scalp maps) where amplitudes
are represented by colors. In such views, the amplitudes between
sensors are interpolated, which allows the entire surface of the head
to be depicted, thus providing a view of the voltage distribution
(i.e., topographies) over the scalp for each experimental condition
as a function of time. As shown in Figure 3.2 visual analysis of the
scalp maps revealed two main focuses of activity in the ERPs: a
bilateral group of electrodes starting at 160 ms and lasting for 200
ms, including 14 electrodes, seven to each side (shown in
discontinuous line). A centro-parietal focus including 14 electrodes
was also found, starting at 247 ms and lasting for 500 ms after
target presentation, which included fourteen electrodes (shown in
continuous line).
Figure 3.2. Group of electrodes selected for the analysis and topographies
showing amplitude differences in non-emotional and emotional tasks.
COMPARING CONFLICT ADAPTATION IN CONFLICT
117
Latencies and mean amplitude voltages averaged over the
selected channels were submitted to repeated-measures analyses of
variance. For the bilateral group of electrodes we conducted a
repeated-measures ANOVA including Hemisphere (left vs right),
Task (non-emotional vs emotional), Previous trial congruency
(congruent vs incongruent) and Current trial congruency (congruent
vs incongruent). For the centro-parietal group of electrodes a
repeated-measures ANOVA with Task (non-emotional vs
emotional), Previous trial congruency (congruent vs incongruent)
and Current trial congruency (congruent vs incongruent) as factors
was carried out.
3. Results.
3.1. Behavioral analysis
Overall accuracy was very high, being error percentage less
than 1% and no difference in accuracy among experimental
conditions was found. Only correct responses were considered in
the RT and the ERPs analyses. RTs shorter than 300 ms and higher
than 1000 ms were excluded as outliers. First trial of each block
was removed in the analysis of sequential effects. To avoid the
influence of repetition priming, we also excluded of the analysis
those trials that were complete repetition (same face, same
expression, same word) of the previous trial (14%).
3.1.1. RT analysis
The ANOVA showed larger RT for the emotional (Mean=
603.75 ms, SD= ±62.3 ms) than for the non-emotional task (Mean=
SE II: EXPERIMENTO 1
118
523 ms, SD= ±48.73 ms), F(1,23)= 147.59, p< .01, and for
incongruent (Mean= 572.25 ms, SD= ±58.93 ms) than for congruent
trials (Mean= 554.5 ms, SD= ±52.34 ms), F(1, 23)= 31.38, p< .01.
As can be seen in Fig. 3.3, we also found a significant
interaction of Task x Congruency F(1,23)= 6.57, p< .05.
Figure 3.3. Behavioral results: interaction effect Task x Current Trial
Congruency
Planned comparisons showed significant congruency effect
in both the non-emotional (F(1, 23)= 25.36, p< .01) and the
emotional (F(1, 23)= 22.24, p< .01) versions of the task, but this
effect was larger in the emotional (congruent Mean = 582, SD =
56.96, incongruent Mean = 615.5, SD = 67.64) than in the non-
emotional (congruent Mean = 517, SD = 47.73, incongruent Mean =
529, SD = 49.73) version.
No significant main effect of the Previous Trial Congruency
variable appeared, and no interaction of this with any other variable
was found either.
COMPARING CONFLICT ADAPTATION IN CONFLICT
119
3.2. Electrophysiological analysis
Separate analyses were carried out on the two different
groups of electrodes shown in Figure 2 namely, the bilateral group
of seven electrodes on each side of the head and the fourteen
electrode central group. For each group we analyzed the latency
and amplitude of the ERP components of our interest. Given the
significant main effect due to the type of task found in our RT data,
latency analyses were particularly appropriate to ascertain whether
or not the difference in time found in the behavioral measure had an
effect on the ERP components.
3.2.1. Bilateral group of electrodes
Figure 3.4 shows the average waves for the two seven
electrode groups corresponding to the left and right hemispheres.
On each side the waves for the congruent and incongruent
conditions are displayed as a function of the Previous congruent
variable. No reference is made to the type of task because, as will
be shown below, no effect of this variable on the analyzed
components was found. In this group of electrodes, our interest
centered on the P1 and N170 components.
SE II: EXPERIMENTO 1
120
Figure 3.4. Average waves of the congruent and incongruent trials when
previous trial was congruent or incongruent for the left hemisphere (in the left)
and right hemisphere (in the right).
3.2.1.1. P1
Voltage analysis for this component was performed on a 20
ms window centered on the first positive peak at 150 ms.
Latency analysis
No main or interaction effects were found in the latency
analysis for this component.
Amplitude analysis
This analysis showed a main effect of Hemisphere, F(1,239=
7.077, p< .05, with larger positive voltage in right electrodes
(Mean= 2.984, SD= ± 2.491) than in left electrodes (Mean= 1.664,
SD= ± 2.010). In addition a close to significance interaction of
COMPARING CONFLICT ADAPTATION IN CONFLICT
121
Previous trial congruency * Current trial congruency was also found
F(1,23)= 3.0255, p= .095.
Z Figure 3.5. Averages representation of the interaction effect Previous Trial
Congruency x Current Trial Congruency
Planned comparisons showed a tendency for the Stroop
effect to be larger when the previous trial was congruent than when
it was incongruent, F(1, 23)= 3.7345, p= .0657 with higher
amplitudes of incongruent trials (Mean= 2.404, SD= ± 2.290)
compared to congruent trials (Mean= 2.256, SD= ± 2.405). No
difference between congruent and incongruent trials appeared when
the previous trial was incongruent. The cognitive vs emotional
nature of the task did not show an effect or interact with any other
factor.
3.2.1.2. N170
We analyzed this component on a 40 ms window centered
on a negative peak at 200 ms.
Latency analysis
Again, no main or interaction effects were found analyzing
the latency of this component.
SE II: EXPERIMENTO 1
122
Amplitude analysis
This analysis showed a significant main effect of
Hemisphere F(1,23)= 5.9336, p< .05. Amplitude of right
hemisphere (Mena= -5.030, SD= ±3.054) was more negative than
amplitude for left hemisphere (Mean= -3.688, SD= ±2.135).
We also found a significant main effect of Previous trial
congruency F(1,23)= 4.5885, p< .05. The amplitude for the
previous incongruent trial (Mean= -4.430, SD= ±2.757) was more
negative than amplitude for the previous congruent trial (Mean= -
4.288, SD= ±2.679). No other significant effect appeared.
3.2.2. Centro-parietal group of electrodes.
Figure 3.6. Average waves of the congruent and incongruent trials for the non-
emotional (A) and emotional (B) tasks and for the previous congruent (left) and
previous incongruent (right) trials.
COMPARING CONFLICT ADAPTATION IN CONFLICT
123
3.2.2.1. N2
For this component we selected a 40 ms window centered on
a posterior negativity peaking at 200 ms.
Latency analysis
No main or interaction effects were found in the latency
analysis.
Amplitude analysis
Amplitude analysis for the N2 component showed a close to
significance main effect of the task variable F(1,23)= 3.2014, p=
.086; the emotional task (Mean= -0.958, SD= ± 2.029) showed less
negativity than the non-emotional task (Mean= -1.430, SD= ±
2.317). Also a close to significance main effect of previous
congruency appeared, F(1,23)= 3.1422, p= .089, with less negativity
for the previous congruent (Mean = -1.123, SD = ± 2.204) than for
the previous incongruent (Mean = -1.265, SD = ± 2.175) condition.
A significant interaction effect between Task x Congruency
was found , F(1, 23)= 6.3945, p<.05
SE II: EXPERIMENTO 1
124
Figure 3.7. Graphic representation of the interaction effect between Task x
Current Trial Congruency.
As can be seen in Fig. 3.7, in the non-emotional task the
difference between congruent (Mean= -1.336, SD= ± 2.398) and
incongruent trials (Mean= -1.524, SD= ± 2.254) was significant
F(1,23)= 4.7750, p< .05 while no significant difference between
congruent (Mean= -1.047, SD= ± 2.106) and incongruent trials
(Mean= -0.869, SD= ± 1.967) was found in the emotional task. The
grater negativity of incongruent trials in the non-emotional task
compared with emotional task was also significant F(1,23) =
5.6048, p< .05.
We also found a significant second order interaction effect
between Task x Previous Trial congruency x Current Trial
Congruency F(1,23)= 6.7584, p< .05 (see Figure 3.8)
COMPARING CONFLICT ADAPTATION IN CONFLICT
125
Figure 3.8. N2 Average amplitude of the congruent and incongruent trials for the
non-emotional (A) and emotional (B) tasks and for the previous congruent (left)
and incongruen trials (right).
Planned comparison showed that in the non-emotional task
incongruent trials (Mean= -1.648, SD= ± 2.199) were more negative
than congruent trials (Mean= -1.312, SD= ± 2.506) when previous
trial was incongruent F(1,23)= 7.598, p<.05, but no difference
between congruent (Mean= -1.361, SD= ± 2.338) and incongruent
trials (Mean= -1.400, SD= ± 2.348) was found when previous trial
was congruent. This pattern of results was different for the
emotional task where congruent trials (Mean= -1.259 , SD= ±2.258)
were more negative than incongruent trials when previous trial was
incongruent (Mean= -0.840, SD= ±1.960) F(1,23)= 14.09, p< .01,
but no differences between congruent (Mean= -0.834, SD= ± 2.179)
and incongruent trials (Mean= -0,898, SD= ± 2.058) appeared when
previous trial was congruent.
3.2.2.2. P3
This component was defined on a 20 ms window centered
on a 420 ms positive peak.
SE II: EXPERIMENTO 1
126
Latency analysis
No significant main or interaction effects were found when
we analyzed the P3 latencies.
Amplitude analysis
Looking at P3 amplitudes we found a significant main effect
of congruency, showing more positive amplitude for congruent
trials (Mean = 4.894 µv, SD = ± 3.052) than for incongruent trials
(Mean = 4.668 µv, SD = ± 3.131), F(1, 23)= 6.036, p< .05. The
analysis also showed a significant second order interaction Task x
Previous Congruency x Congruency F(1, 23)= 5,14, p< .05. (see Fig
3.9)
Figure 3.9. P3 average amplitude of the congruent and incongruent trials for the
non-emotional (A) and emotional (B) tasks and for the previous congruent (left)
and previous incongruent (right) trials.
In the non-emotional task, planned comparisons showed
significantly less amplitude for incongruent trials (Mean= 4.26;
SD= ± 2.948) compared with congruent trials (Mean= 4.801, SD= ±
COMPARING CONFLICT ADAPTATION IN CONFLICT
127
3.081) when they were preceded by an incongruent trial F(1,23)=
11.12, p< .01. Also the difference in amplitude between incongruent
trials following an incongruent trial and those following a congruent
one was significant F(1,23)= 4.5712, p< .05. No other significant
difference was found either in the non emotional or the emotional
tasks.
4. Discussion
The main purpose of the present experiment was to further
understand the relationship between emotional and cognitive
conflict using a face-word Stroop task where both, emotional as
well as cognitive conflict, were included in the same task. By
recording the scalp electrical activity of the subjects while they
were performing the task we also tried to study the brain processes
associated with the two types of conflicts.
Our RT data showed that the task employed was adequate to
create conflict situations so that responses during incongruent trials
took longer than responses during congruent trials in both the
cognitive and emotional versions of the task. Our results for the
cognitive conflict situation replicates the familiar Stroop effect
found in a wide series of experiments (see Mac Leod, 1991). Also
our emotional conflict results support previous research using an
emotional Stroop task (vg. Mckenna & Sharma, 1995; Etkin, Egner,
Peraza, Kandel & Hirsch, 2006) or emotional counting Stroop task
(Bush., Whalen., Rosen., Jenike., McInerney., & Rauch., 1998;
Hayward., Goodwin., & Harmer., 2004), though, as said above, in
these experiments emotional information acted as distracter rather
SE II: EXPERIMENTO 1
128
than target. We used emotional information as target and examined
emotional and cognitive conflict jointly in the same experiment. In
agreement with similar previous research (Alguacil et al., 2013;
Egner et al., 2008) we found a significant Stroop effect for both, the
non-emotional and the emotional versions of the task, though this
effect was higher in the emotional than cognitive version. One
possible explanation for this difference may be related to the fact
that the RT of the emotional version was significantly longer than
the RT of the cognitive version of the task. It appears that
identifying a male or a female face is relatively easy compared with
discriminating the emotional expression of those faces; while
identifying the face gender may be based on global characteristics
of the face, discriminating its emotional expression may require to
focus attention on specific facial features like eyes, nose and mouth.
In fact, influential models of face perception (Bruce & Young,
1986; Haxby & Gobini, 2011) consider gender, as a permanent
characteristic, to be part of a face identity (core system according to
Haxby & Gobini, 2011) that is processed at early stages of face
perception. However emotional facial expression, being a
changeable characteristic, requires additional information and is
processed at later stages of face perception (extended system for
Haxby and Gobini, 2011). This line of reasoning is further
supported by the behavioral results reported by Egner et al. (2008);
using a task similar to ours, also found larger RT for the emotional
than for the non-emotional task. On the other hand, Alguacil et al.
(2013) found no difference between the two types of tasks, using
words instead of faces. Therefore it is likely that the difference in
COMPARING CONFLICT ADAPTATION IN CONFLICT
129
our data may be due to the type of stimuli on which the response
was made. In any case, if the emotional version of the task turns out
to be more difficult, as it is the case in the present experiment, is not
surprising that it also gives way to a greater Stroop effect compared
to the cognitive version of the task.
Our RT data did not show any effect due to the Previous trial
congruency variable, though this variable did influence our ERP
data. Why conflict adaptation was not present in our RT data is
difficult to explain. Alguacil et al. (2013) and Oschner et al. (2008)
did not find conflict adaptation in their behavioral data either.
Alguacil et al. (2013) pointed to the complexity of the task as a
possible explanation, but Egner et al. (2007), using faces in a task
similar to ours, reported a conflict adaptation effect in both, the
emotional and non-emotional tasks. Further research is needed on
this issue.
Despite the large difference in RT between the emotional
and cognitive tasks, in our ERP data no component in either group
of electrodes showed differences in latency due to the emotional or
cognitive characteristics of the task. It appears that the difference
found in the RT data may be due to differences in later components
related either to response preparation, to response execution or both.
Nevertheless, neither the latency of the early components, P1 and
N170, nor that of the components mainly related to conflict, N2 and
P3, changed as a function of the emotional or cognitive nature of
the task.
SE II: EXPERIMENTO 1
130
In the posterior bilateral set of electrodes, the P1 component
showed a significant difference in amplitude depending upon the
laterality of the electrode location; the amplitude of electrodes on
the right side was larger than the amplitude of electrodes on the left
side of the head. This result is in agreement with the well-known
cerebral dominance of the right hemisphere for face processing (see
Kanwisher and Barton, 2011). In addition, our data on P1 provide
some indication of conflict adaptation, since the Stroop effect
decreased when previous trial was incongruent relative to previous
congruent trial. However, this tendency was only close to
significance.
The amplitude of the N170 component replicated the
findings on P1 concerning laterality; voltage values were more
negative for electrodes on the right than for electrodes on the left
side of the head. This result is in agreement with to the view that
the right lateralized N170 may be considered an
electrophysiological marker of face processing (Eimer, 2011). In
the present experiment, the N170 also appeared sensitive to
Previous congruency, as the amplitude were more negative for
previous incongruent than for previous congruent trials. However
no sign of the interaction between Previous and Current
congruency, and consequently of conflict adaptation, appeared on
this component.
Taken together, the results of the this experiment did not
replicate the clear-cut findings of Alguacil et al. (2013) with respect
to the early, P1 and N170, components. These authors found
COMPARING CONFLICT ADAPTATION IN CONFLICT
131
conflict adaptation effects for both tasks on these two early
components and, following conflict monitoring theory (Botvinick et
al., 2001), related the effect on P1 to conflict detection and the
effect on N170 to control implementation. The main difference
between our procedure and that of Alguacil et al. lies on the type of
stimuli used, as they employed words instead of faces. It is
difficult, however, to ascertain how this difference may have
generated different results.
In the present experiment the N2 component of interest had
a central and posterior topography and appeared relatively early,
picking around 200 ms. This posterior N2, also called N2c
(Pritchard, Shappell and Brant, 1991), has been related to stimulus
classification rather than cognitive control (Folstein and Van
Petten, 2008). However in our experiment the posterior N2 was
clearly modulated by conflict resolution and previous congruency.
In the cognitive version of the task the N2 displayed negative
amplitude larger for incongruent than congruent trials in agreement
with previous results showing N2 modulations when attention is
focused on targets avoiding irrelevant information (Luck, 2005).
These modulations appear related to the response, by monitoring the
inappropriate responses that in conflicting situations may compete
with the correct one (Folstein and Van Petten, 2008). Also in the
cognitive task we found a significant modulation of this component
as a function of previous congruency, so that the more negative
amplitude for incongruent compared with congruent trials was
largest when they followed an incongruent trial. In contrast,
conflict adaptation showed a different pattern in our emotional task
SE II: EXPERIMENTO 1
132
since, after an incongruent trial, the N2 amplitude was significantly
shorter for incongruent than for congruent trials. These results
show a clear dissociation on N2 conflict adaptation between the
emotional and the non-emotional tasks.
There is no agreement regarding the way that conflict
adaptation should be reflected on N2 in cognitive tasks. Clayson
and Larson (2011) using a modified Eriksen flanker task, in which
participants had to identify the direction of a central arrow pointing
to the right or to the left, reported a decrease of N2 for incongruent
trials following another incongruent trial. On the other hand, we
found more negative amplitude for incongruent compared with
congruent trials when they followed an incongruent trial. There is a
difference in difficulty between our task and the Eriksen flanker
tasks, employed by Clayson and Larson, that may explain the
diversity of results. If we assume that the N2 component is related
to focusing attention on targets and avoiding irrelevant information
(Luck, 2005), responding to the gender or emotion of a face appears
more attention demanding than responding to the direction of a
pointing arrow. It is reasonable to assume that the amount of
attention focusing induced by an incongruent trial on the following
trial may depend upon the difficulty of the task, being greater for
high demanding tasks. As a result, an incongruent trial following
another incongruent trial may receive more attention, shown in a
greater N2, than a congruent trial in a high demanding task while it
may not need additional attention in a low demanding task. Though
this explanation is clearly ad hoc, it provides an interesting
COMPARING CONFLICT ADAPTATION IN CONFLICT
133
prediction about conflict adaptation being dependent upon the
difficulty of the task, at least in cognitive tasks.
Concerning the emotional version of the task, our results
showed N2 amplitude significantly shorter for incongruent than for
congruent trials after an incongruent trial. This is likely the first
time that conflict adaptation on N2 is reported in an emotional
Stroop task in which participants had to perform explicit affective
evaluations on the stimuli; in fact, Alguacil et al. (2013) did not
observe any N2 modulation in their affective task when they asked
their subject to evaluate the emotional meaning of words. As
mentioned above, experiments showing emotional conflict on N2
have generally asked their subjects to respond according to the
cognitive nature of target stimuli displayed upon an emotional
background (Kanske & Kotz, 2010, 2011).
It is not easy to explain why in our experiment emotional
conflict adaptation on N2 appears as amplitude shorter for
incongruent than for congruent trials. According to the conflict-
monitoring hypothesis we would expect a control reduction in an
incongruent relative to a congruent trial as a consequence of being
preceded by another incongruent trial. If N2 amplitude were
directly related to control, then our data would lend support to a
reduction of control for incongruent compared with congruent trials
after an incongruent trial, but only in our emotional task, not in the
cognitive task. Thus, it appears that the conflict adaptation
hypothesis cannot explain our results for both, our cognitive and
emotional task. It is also possible that our results provide
SE II: EXPERIMENTO 1
134
supporting evidence for neuronal mechanisms different for
cognitive than for emotional control; Egner et al. (2008), using
fMRI and the same task as we have used, reported two distinct
neuronal control circuits: a lateral prefrontal system, in charge of
resolving cognitive conflict, that was associated with enhance
processing of task relevant stimuli, and a rostral anterior cingulate
system, resolving emotional conflict, and associated with decreased
amygdalar responses to emotional distracters. These two
mechanisms could explain our N2 results for both, the cognitive and
emotional tasks. In the cognitive task, the N2 amplitude larger for
incongruent than congruent trials, would be mainly related to the
working of the lateral prefrontal system, enhancing task relevant
stimuli. In the emotional task, the N2 amplitude, shorter for
incongruent than congruent trials, would mainly reflect a decreased
in amygdalar responses to the emotion of the face. Further research
is needed to replicate our findings and to ascertain the different
factors that may influence N2 amplitude in emotional and non-
emotional conflict situations. In a relevant review on cognitive
control and the N2 component, it was concluded that even the N2c
may need subdivisions (Folstein and Van Petten, 2008). If, in
addition, emotional control is taken in account, the need for
research on N2 is further increased.
With reference to the P3 component, we found again a
second order interaction between Previous congruency and
Congruency that was different for the non-emotional than for the
emotional task. A significant conflict adaptation effect appeared in
the non-emotional task but no significant difference was found in
COMPARING CONFLICT ADAPTATION IN CONFLICT
135
the emotional task. In the non-emotional version, conflict
adaptation appeared as a decrease in P3 amplitude of incongruent
trials following an incongruent trial compared with incongruent
following a congruent trials. These results replicate those reported
by Clayson and Larson (2011). However Alguacil et al. (2013) did
not find conflict adaptation in any of the three different windows
defined on P3.
It has been suggested that the P3 component in conflict
situations is related to control operations such as the allocation of
attentional resources and inhibition of conflicting responses (Polich,
2006). Our P3 results in the non-emotional task lend support to the
conflict-monitoring hypothesis that would predict a reduction in P3
amplitude for incongruent trials following an incongruent trial
compared to incongruent following congruent trials. According to
this hypothesis, conflict adaptation results from activation of the
DLPFC that serve to prevent conflict in subsequent performance.
Thus, when an incongruent trial is presented, dACC detects the
conflict situation triggering signals to DLPFC that implements
strategies to resolve this situation. If a subsequent incongruent trial
is presented, the DLPFC is now prepared for the resolution of the
conflict, investing less attentional resources than in the previous
incongruent trial. Therefore our data also support a possible
relationship between P3 amplitude and the control activity of the
DLPFC, at least in non-emotional conflict situations. Egner et al.
(2008) found a relation between DLPFC activation and the
resolution of conflict, cognitive as well as emotional. However, in
our emotional task we were unable to find any differences on P3.
SE II: EXPERIMENTO 1
136
As a consequence, it is likely that the relationship between DLPFC
activity and P3 amplitude is multifaceted and need further research.
In sum, the present experiment provides evidence of conflict
adaptation on ERP components closely related to control namely,
the N2 and P3 components. To our knowledge, this experiment and
the one by Alguacil et al. (2013) are the only ones conducting a
direct comparison of the time course of cognitive and emotional
conflict processing using EEG recordings in a task including an
explicit evaluation of both types of materials. Curiously enough,
Alguacil et al. found scalp signatures of conflict adaptation on the
early P1 and N170 components, where we were unable to find
them. In contrast, we found such scalp signatures on N2 and P3,
where Alguacil et al. (2013) did not. The main difference between
the two experiments have to do with the task and kind of stimuli
employed; Alguacil et al., used an Eriksen flanker task with only
words as stimuli, while we have utilized a Stroop task with words
and faces. Further research will hopefully shed light upon these
discrepancies.
Capítulo 4
SERIE EXPERIMENTAL II:
Unconscious CSPCE in cognitive and
emotional conflict: An ERPs study
1. Introduction
Cognitive control processes have been a central topic in
cognitive neuroscience research. As in real life, many experimental
tasks require selection of relevant information to achieve an
accurate performance. Cognitive control plays a role when the task
environment contains both relevant and irrelevant task information.
Stroop task (Stroop, 1935 experiment 2) provides an example of
this situation. In this task participants have to name the ink color in
which color words are printed. Two different conditions are usually
presented: congruent and incongruent. In the congruent condition,
the ink color matches the color name of the word (e.g., RED in red
ink). In the incongruent condition, by contrast, the ink color is
different from the color name of the word (e.g., BLUE in red ink).
To perform the task in the incongruent condition, participants, by
using control strategies, must avoid the automatic process of
reading the word and have to name the color of the ink in which the
word is printed. In the congruent condition, the ink color matches
the color name of the word, so avoidance of the automatic reading
process is not necessary. Consequently, reaction times (RTs) are
longer in the incongruent condition than in the congruent one. This
cost of time is known as Stroop interference.
Stroop interference can be affected by the relative proportion
of congruent and incongruent trials (Logan & Zbrodoff, 1979; Lowe
& Mitterer, 1982; West & Baylis, 1998). That is, when the
proportion of incongruent trials is high, the Stroop effect is smaller
than in situations in which the proportion of congruent trials is
SE II:EXPERIMENTO 2
140
predominant. It seems as if frequent experience with conflicting
stimuli or response features facilitates the resolution of interference.
In general terms, a kind of facilitation like this is known as a
Conflict adaptation effect (CAE) (Botvinick, Braver, Barch, Carter,
& Cohen, 2001). Several factors responsible of this effect have
been advanced. The main ones are the overall proportion of
congruency at the list level (proportion of congruency effect, PCE)
(Logan, Zbrodoff, & Williamson, 1984) or a more specific and
online processing of the information either at the item level (item-
specific proportion congruency effect, ISPCE) (Jacoby, Lindsay and
Hessels 2003) or at the context level (context-specific proportion
congruency effect, CSPCE) (Bugg, Jacoby, & Toth, 2008; Crump,
Vaquero, & Milliken, 2008; for a review, see Bugg & Crump,
2012).
Proponents of the PCE (Logan, et al., 1984) argue that in a
Stroop task subjects become aware of the contingency between the
color and the word and adapt their strategies to attend the word
when the proportion of congruent trials is higher than that of
incongruent ones and ignore the word when the proportion of
congruent and incongruent trials is reversed. Consequently, the
Stroop effect is larger when the proportion of congruent trials is
predominant because the color word is likely to be attended to in
incongruent trials as well. By contrast, when the proportion of
congruent trials is lower than that of incongruent trials, the Stroop
effect is smaller than in the previous case because the color word is
mostly ignored in incongruent trials. According to this approach,
the proportion of congruent items influences performance at a list-
UNCONSCIOUS CSPCE IN CONFLICT & ERPs
141
wide level by switching participants’ attention to one of the two
dimensions depending on which one is relevant for the task.
Proponents of the ISPCE contend that the proportion
congruency effect acts at the item level. Jacoby et al. (2003) chose
six colors and their corresponding color words and divided them
into two sets of equal size. To produce mostly congruent (MC)
items, words in one set were presented in their congruent colors in
eighty percent of trials and in another color from that set in the
remaining twenty percent of trials. These rates were reversed in
words from the other color set to produce mostly incongruent (MI)
items. The proportion of congruency at the list level was fifty
percent. This manipulation of proportions prevents subjects from
developing a general list-level strategy because the proportion of
congruent and incongruent items is the same. However, Jacoby et
al. found that MC items showed a larger Stroop effect than MI
items. Results similar to those of Jacoby et al. were obtained by
Crump, Gong and Milliken (2006) in a situation in which the
likelihood of congruency was associated to a task-irrelevant
location context rather than to a particular color word. According to
these authors, the ISPCE belongs to a larger class of effects known
as context-specific proportion congruency effect (CSPCE) that is
driven by the relationship between the context and the likelihood of
congruency. In recent years, the CSPCE has been reported in a
range of Stroop-like effects (Bugg, Jacoby, & Toth, 2008; Crump,
Vaquero, & Milliken, 2008; for a review, see Bugg & Crump, 2012)
and in different types of contexts, including social categories
(Cañadas, Rodríguez-Bailón, Milliken, & Lupiáñez, 2012).
SE II:EXPERIMENTO 2
142
A second example of conflict adaptation has to do with
sequential effects. Gratton, Coles and Donchin (1992) first studied
these effects using the Eriksen flanker task. They found that RT
decreased for current incongruent trials when they followed another
incongruent trial compared with when they followed a congruent
trial. On the other hand RT increased for congruent following
incongruent trials compared with congruent following congruent
ones. This result, known as Gratton effect, has been found across
different conflict tasks, including Stroop (e.g., Egner & Hirsch,
2005), Simon (Hommel, Proctor & Vu, 2004) and flanker tasks
(e.g., Nieuwenhuis, Stins, Posthuma, Polderman, Boomsma & De
Geus, 2006); it seems to be domain-specific, as it does not transfer
across different tasks that are performed sequentially (Egner, 2008;
Funes, Lupiáñez, & Humphreys, 2010). The most generally
accepted explanation for this effect is that participants detect
conflict on incongruent trials and decrease attention to the word on
the following trial in order to avoid further conflict. As a result, the
Stroop effect will be smaller. In contrast, on congruent trials there is
no conflict, so attention will not be as constrained on the following
trial. Hence, the word can interfere more strongly and the Stroop
effect will be larger. Due to these processes, a Gratton effect will
emerge, that is, an interaction between congruency on the current
trial and congruency on the previous trial.
At computational and neural levels of analysis, Botvinick et
al. (2001); (see also Botvinick, Cohen, & Carter, 2004; MacDonald,
2000) advanced a common explanation for both, the PCE and the
Gratton effects, in what they termed the conflict-monitoring
UNCONSCIOUS CSPCE IN CONFLICT & ERPs
143
hypothesis. According to these authors, the Stroop task is just a
particular case of conflict resolved by the interplay of a set of neural
structures. The dorsal anterior cingulate cortex (dACC) acts as a
detection mechanism that responds to the occurrence of a conflict
situation (i.e., an incongruent trial in a Stroop task). The conflict
signal triggers strategic adjustments in the dorsolateral prefrontal
cortex (DLPFC) and other posterior and subcortical structures that
serve to prevent conflict in subsequent performance. Accordingly,
in a list where incongruent trials outnumber congruent ones,
strategic adjustment mechanisms will be highly operative, thus
minimizing the influence of the color word on performance and
reducing the Stroop effect. For the Gratton effect the explanation is
similar. When an incongruent trial is presented, dACC detects the
conflict situation triggering signals to DLPFC that implements
strategies to resolve this situation. If a subsequent incongruent trial
is presented, the DLPFC is now prepared for the resolution of the
conflict, which translates in faster RT in the current incongruent
trial. Experiments using functional magnetic resonance (fMRI) have
lent support to this conflict-monitoring hypothesis (Botvinick,
Cohen, & Carter, 2004; Kerns, Cohen, McDonald III, Cho, Stenger
& Carter, 2004; McDonald, 2000). Nevertheless, from a
computational approach, Blais, Robidoux, Risko and Besner, (2007)
showed that Botvinick’s conflict-monitoring model could not
explain the ISPCE and proposed a modified model according to
which the DLPFC exerts control at the specific item level rather
than at the general list level. Likewise, in a recent experiment using
fMRI, Blais and Bunge (2010) found that many of the regions
SE II:EXPERIMENTO 2
144
involved in cognitive control in the Stroop task, including the ACC
and DLPFC, were operative at a local item level.
Conflict not only exists in the cognitive but also in the
emotional domain. Many tasks have been adapted in order to study
the conscious emotional conflict resolution processes (Bishop,
Duncan, Brett, & Lawrence, 2004; Bush, Luu, & Posner, 2000;
Compton, 2003; Shin, Whalen, Pitman, Bush, Macklin, Lasko et al.,
2001; Whalen, Bush, Mcnally, Wilhem, McInerney, Jenike &
Rauch, 1998). For example, in the emotional counting Stroop task
(Whalen et al., 1998), participants have to report the number of
words that appear on a screen, regardless of the word meaning.
Control trials contain emotionally neutral words (e.g., ‘cabinet’
written three times), while interference trials contain emotional
words (e.g., ‘murder’ written three times). As a general finding,
emotional words tend to show greater interference effect (longer
RT) when compared with neutral (non-emotional) words. Similar
results have been found in emotional Stroop tasks where subjects
have to name the ink color of words the meaning of which may be
emotional or non-emotional. As in the previous task, emotional
words produce longer RT than non-emotional words (e.g. Watts,
McKenna, Sharrock & Trezise, 1986).
A common characteristic to these tasks is that emotional
information acts as distracter rather than target, because targets are
always cognitive. Thus they may be suited to provide information
about the interference produced on the control of a cognitive
activity (either counting or color naming) by emotional information,
UNCONSCIOUS CSPCE IN CONFLICT & ERPs
145
but they do not provide direct information on emotional control. To
achieve this information, emotional stimuli should be included as
targets. Recently, attempts have been made to study both, cognitive
and emotional, control mechanisms within the same task in order to
equate the cognitive and affective type of conflict. The face-word
Stroop task (Egner, Etkin, Gale, & Hirsch, 2008; Etkin, Egner,
Peraza, Kandel, & Hirsch, 2006, Panadero & Tudela, submitted)
may be considered a good example. In the emotional version of this
task, subjects have to name the emotion displayed by either a male
or female face while an emotional word is presented across the face
such that the meaning of the word and the emotion of the face may
be either congruent or incongruent. In the cognitive version of the
task, subjects have to respond to the gender (male or female) of the
faces disregarding the meaning of the word (male or female) printed
across the face that may be congruent or incongruent with the face
gender. Results show that in both versions of the task the
incongruent condition creates conflict (emotional or cognitive),
reflected in slower RT in incongruent relative to congruent trials.
Other researchers (Alguacil, Tudela, & Ruz, 2013; Ochsner,
Hughes, Robertson, Cooper, & Gabrieli, 2008) have employed a
variant of the Eriksen flanker task (Eriksen & Eriksen, 1974) in
which a word was presented as target at the center of the screen and
flanked by another two words above and below. In the emotional
version, participants had to respond to the valence of the target
(positive or negative); in the cognitive version they had to indicate
the target’s semantic category. In either version, the flanker words
could be congruent or incongruent with the target. As in the Stroop
SE II:EXPERIMENTO 2
146
task, also in the flanker task the incongruent trials create both,
cognitive and emotional conflict (slower RT).
Regarding neural mechanisms underlying cognitive and
affective conflict, an influential review of experiments, using
imaging techniques, on cognitive and emotional versions of the
Stroop task (Bush, Luu, & Posner, 2000) suggested a segregation of
functions within the anterior cingulated cortex (ACC); the dorsal
part would be mainly involved in cognitive control while the ventral
ACC (vACC) would be engaged when emotional control is needed.
This hypothesis was based in experimental evidence where
cognitive and emotional control was studied in experiments in
which, as pointed out above, emotional information acted as
distracter for cognitive responses. More recently, the neural
substrate of cognitive and affective conflict has been jointly
explored within the same experiment using fMRI (Egner, et al.,
2008; Ochsner, et al., 2008). Their results partially modified Bush et
al.’ hypothesis so that emotional conflict seems to recruit vACC and
rostral medial prefrontal cortex (RMPC) while cognitive conflict
appears under control of the DLPFC.
To date, few experiments have studied the ERPs associated
to cognitive and emotional conflict jointly in the same experiment
and using the same kind of material and task settings across
conditions (Alguacil, et al., 2013, Panadero & Tudela, to be
submitted). In the Alguacil et al. experiment, participants performed
a flanker task in which they had to focus on a central target and
indicate its semantic category (cognitive version) or its valence
UNCONSCIOUS CSPCE IN CONFLICT & ERPs
147
(affective version). Targets were flanked by congruent or
incongruent words in both versions. Results showed differences in
the wave amplitude as a function of the cognitive or emotional
nature of the task in the late ERP components but none in the early
components. They found conflict adaptation effects in N170
component with larger amplitudes for incongruent trials preceded
by a previous incongruent trial. This pattern was the same for both,
emotional and non-emotional task, showing that conflict had been
processed in a very early stage of information processing. They also
found that the N2 and the first part of the P3 were exclusively
modulated by cognitive conflict, whereas the last section of the P3
deflection was only involved in affective conflict processing. The
authors argued in favor of the existence of early common
mechanisms, equivalent for cognitive and affective materials, and
later task-specific conflict processing.
In contrast, Panadero and Tudela (to be submitted), using
faces instead of words as stimuli, did not find conflict adaptation in
the N170 potential, but did find it in both, the N2 and P3 potentials;
in the N2 potential there was a dissociation between the cognitive
and the emotional tasks; in the former, the N2 component for
incongruent trials increased in negativity when they were preceded
by another incongruent trial, while in the latter, the N2 component
decreased in negativity under the same circumstances. In addition,
they found evidence of conflict adaptation in the P3 component
only for the cognitive, but not for the emotional, task. Panadero and
Tudela argued that their results might provide electrophysiological
support for neural mechanisms of control different for non-
SE II:EXPERIMENTO 2
148
emotional and emotional situations. In the present experiment we
tried to further explore the relationship between cognitive and
emotional conflict and their underlying neural mechanisms.
A common characteristic of the experiments mentioned so
far, is that they have studied conflict, either cognitive or emotional,
in situations where participants were well aware of the stimuli
presented and most of the relationships among them. As a result,
theories such as the conflict monitoring hypothesis understand
control as a conscious and strategic process. Few experiments have
studied the possible unconscious nature of conflict adaptation
processes. Klapp (2007, Experiment 3) explored the PCE under
masked conditions using a masked arrow paradigm. He used a
spatial Stroop-like task involving successively presented arrows
pointing in either the same (compatible condition) or opposite
direction (incompatible condition). The first arrow was presented
for 32 ms and immediately followed by the mask together with the
second arrow so that the first arrow was unnoticed to the subjects.
The results yield a significant interaction so that the difference in
RT between compatible and incompatible conditions was higher
when the arrows pointed in the same rather than when they pointed
in the opposite directions. However, in Klapp’s experiment (2007)
participants received feedback for incorrect responses, which makes
it difficult to interpret his results. It may be that subjects’ responses
adapted to the conscious error rates rather than to the unconscious
frequency of congruent and incongruent trials. Heinemann, Kunde
and Kiesel (2009) explored the CSPCE under masked conditions
using a task in which subjects had to categorize target numbers as
UNCONSCIOUS CSPCE IN CONFLICT & ERPs
149
being larger or smaller than five. A masked number (i.e., the prime),
that could be congruent or incongruent with the target, preceded the
target. At the beginning of each trial, a colored rectangle was
presented as background simultaneously with the fixation cross.
The color of the rectangle was associated with a particular
congruency context that could be either 80% or 20%. Results
showed that the difference in RTs between congruent and
incongruent trials was significantly higher in the 80% than in the
20% congruent trial condition but only when subjects were able to
see the prime; no differences were observed when they could not
see it. (Heinemann et al., 2009). The authors concluded that the
CSPCE requires conscious representation of the conflicting
information, namely prime, target and context (Kunde, Reuss, &
Kiesel, 2012).
Recently, Panadero, Castellanos and Tudela (2015) reported
evidence for the unconscious nature of the conflict adaptation
effect, specifically the context-specific proportion congruency effect
(CSPCE). Following in part a procedure introduced by Blais,
Tudela and Bunge (2007), they used a masked Stroop-like task
where two different proportions of congruency were associated to
two dissimilar masks. In addition they used electrophysiological
measures in an attempt to trace the time course of each trial neural
processing. Their results provided clear evidence of unconscious
CSPCE not only in the RT but in the N2 and P3 components as
well; a significant interaction between the percentage of congruency
and the congruency between the color word and the color of the
target, appeared in the three dependent variables, showing less
SE II:EXPERIMENTO 2
150
Stroop effect when the proportion of incongruent trials associated
with one of the masks was 66%, than when the proportion of
incongruent trials associated with the other mask was 33%.
In relation to emotional conflict, to our knowledge, no
evidence of unconscious CSPCE has been provided so far. Thus,
the main purpose of the present experiment was to study both,
unconscious cognitive and emotional conflict, and in particular the
possible existence of unconscious CSPCE in emotional conflict, in a
task similar to that employed by Panadero and Tudela (submitted).
We used a face-categorization task (Egner y Hirsch, 2005b; Gale;
and Hirsch, 2007), in which we varied the source of conflict
between non-emotional and emotional stimulus representations,
while keeping task-relevant stimulus characteristics constant. The
task-relevant information for both tasks was a set of male and
female faces expressing either happiness or fear. For the cognitive
task, subjects had to respond to the gender of the face while
ignoring gender labels appearing across the faces. For the emotional
task subjects had now to respond to the emotion expressed by the
face ignoring a congruent or incongruent emotional label across the
faces. To insure the unconscious processing of the labels, we
followed a procedure similar to that employed by Panadero,
Castellanos and Tudela (2015). Thus, the labels were preceded and
followed by masks in both, the cognitive and emotional, tasks. In
turn, each mask was associated to a particular proportion of
incongruent trials. During the entire experiment we recorded
continuously the EEG of the participants.
UNCONSCIOUS CSPCE IN CONFLICT & ERPs
151
In summary, we tried to explore the emotional and non-
emotional conflict effects under unconscious conditions and to
observe how the proportion of congruency modulates these effects
when it is associated to a particular context. We also took into
account the effect of previous trial congruency to see how the
CSPCE and the Gratton effects might interact. We focused on
analyzing the unconscious situation but we tried to maintain an
expectation of the presence of a prime word between masks by
calibrating the duration of this word individually for each
participant until it was unnoticed. It has been recently reported
(Martens, Ansorge, & Kiefer, 2011) that subliminal semantic
priming can be modulated by attentional task set. In this experiment
we tried to keep the task set constant while the processing of the
prime word remained unconscious.
2. Material and methods
2.1. Participants
Twenty-four students of the University of Granada (23
females and 1 male) participated in a one-session experiment. A
group of twelve participants conducted a non-emotional version of
the task and another group of twelve participants conducted an
emotional version of the same task. Participants reported having
normal or corrected-to-normal vision and were not aware of the
purpose of the experiment. Mean age was 21.67 years (SD = ±2.37).
Participants received course credit in exchange for their
participation and signed a consent form approved by the local
Ethics Committee.
SE II:EXPERIMENTO 2
152
2.2. Stimuli and Apparatus:
Three different types of stimuli were presented: twelve face
pictures (6 men and 6 women, three of them were happy and three
were angry), two different masks and two kinds of words. For the
non-emotional version the words were MASCULINO (4.581º wide
and 0.573º high) and FEMENINO (3.8º wide and 0.573º high). In
turn, for the emotional version the words were ASUSTADO (3.347º
wide and 0.573º high) and CONTENTO (3.347º wide and 0.573º
high). Masks could be either @@@@@@@@@ (4,581º wide and
0,802º high of visual angle) or ######### (6,867º wide and 0,802º
high of visual angle), each one associated with a different
percentage of incongruency and, counterbalanced per participants;
for odd participants @@@@@@@@@ was associated to a 66%
and ######### to a 33% incongruency percentage. This was
reversed for even participants. Words and masks were presented in
18 Courier New letter in white color on a black background.
Targets were face pictures of about 10.28º wide and 11.42º
high of visual angle; they appeared in the center of the screen. All
stimuli were presented on a 17 inch color screen of a PC computer
at a viewing distance of about 50 centimeters. The PC was
connected to a Macintosh computer for the registration of the ERPs.
The entire task was created and displayed using the E-Prime 1.2
Professional software (Schneider, Eschman & Zuccolotto, 2002).
UNCONSCIOUS CSPCE IN CONFLICT & ERPs
153
2.3. Procedure and design
2.3.1 Procedure
The experimental session had two parts: a threshold setting,
followed by an experimental section. In the first part, the awareness
threshold for the words was determined for each participant in the
following way. First a cross served as a fixation point and was
presented in the center of the screen for 1000 ms. Then a word that
could be ASUSTADO or CONTENTO for the emotional task and
MASCULINO or FEMENINO for the non-emotional task, appeared
immediately preceded and followed by the same mask. The
preceding mask lasted for 109 ms and the following 689 ms. At the
beginning the word was presented for 100 ms, a time interval that
participants could easily perceive. They had to report whether or not
they were able to see the word. If the answer was affirmative we
decreased the presentation time of the word in 10 ms steps
(equivalent to one refresh rate of the computer screen). If it was
negative, we kept the presentation time of the word constant and
repeated the procedure until the answer remained negative for three
consecutive trials. The obtained presentation time was kept as the
awareness time threshold for the particular subject. The second part
of the session consisted of eight masked blocks with seventy-two
trial each with the following event sequence: A fixation point
appeared during 1000 ms and was followed by the first mask lasting
109 ms. Then one word was presented for the time determined in
the previous part of the experimental session and followed by the
same mask now lasting 689 ms. The SOA value between the second
SE II:EXPERIMENTO 2
154
presentation of the mask and the target stimuli was 1000 ms. These
target stimuli were face pictures of scared or happy men or women
and were displayed during 1500 ms. Participants responded by
pressing z or n in a Qwerty keyboard. Keys were counterbalance per
participant. The subject response was followed by an inter-trial
interval that varied randomly between 1000 and 1500 ms (see Fig.
4.1.A).
In the non-emotional task participants had to respond to the
gender of the face and in the emotional task they had to respond to
the emotion of the faces. At the beginning of the entire procedure
we instructed the participants to concentrate on the task and try to
respond as fast as possible to the target. After the experimental
session every participant had to respond to three questions to ensure
they were not aware of the association between the incongruency
percentage and the mask.
2.3.2. Design
Within each block of 72 masked trials three within-subject
independent variables were manipulated: Incongruency percentage
(33% vs 66%), Previous trial congruency (congruent vs
incongruent) and Current trial congruency (congruent vs
incongruent). In the 33% incongruency mask condition, 24 trials
were congruent and 12 incongruent. In the 66% incongruency mask
condition, 12 trials were congruent and 24 incongruent. Thus
within each block the number of congruent and incongruent trials
remained the same and each mask appeared an equal number of
times but associated with a different percentage of incongruency.
UNCONSCIOUS CSPCE IN CONFLICT & ERPs
155
Finnally, an additional between-subject variable was manipulated:
Task (emotional vs non-emotional).
2.4. EEG recording and data analysis:
Participants were seated in front of the computer monitor in
an electrically shielded room. They were instructed to avoid eye
blinks and movements during target presentation and response.
EEG was recorded using a high-density 128-channel Geodesic
Sensor Net (Tucker et al., 1994). The head coverage included
sensors lateral to and below both eyes to monitor horizontal and
vertical eye movements. The EEG net was connected to an AC-
coupled, high-input impedance amplifier (200 M). At the
beginning of the recording session, impedance for each channel was
measured and kept under 50 kΩ, as recommended for Electrical
Geodesics high-input impedance amplifiers. Amplified analog
voltages (0.1–100 Hz band pass) were digitized at 250 Hz (12 bits
A/D converter and 0.02 V minimum resolvable voltage). All
channels were referenced to the Cz electrode during the recording
and were algebraically re-referenced off-line to the average.
The continuous EEG was filtered offline using a 30 Hz low-
pass filter. After that it was segmented in epochs of 200 ms before
and 600 ms after target onset. A 200 ms segment previous to the
target presentation was used to calculate the baseline. The epochs
were submitted to software processing for identification of artifacts.
Segments with eye movements, eye-blinks (electro-oculogram
channel differences greater than 70 V), more than 20% of bad
channels or corresponding to incorrect responses were not included
SE II:EXPERIMENTO 2
156
in the ERPs. Data from consistent bad channels were later replaced
using a spherical interpolation algorithm (Perrin, Pernier, Bertrand,
& Echallier, 1989). A minimum criterion of 30 artifact-free trials
per participant and condition was established to maintain an
acceptable signal-to-noise ratio. A final grand-average was obtained
per condition by pooling the subject’s averages in each
experimental condition.
To facilitate the selection of spatio-temporal windows for
amplitude analyses, the topographic maps provided by the Net
Station Viewer (EGI, 2008) were used as a guide. A topographic
map view displays samples as representations of the scalp projected
onto disk-shaped maps (also referred to as scalp maps) where
amplitudes are represented by colours. However, in the topographic
map view the amplitudes between sensors are interpolated, which
allows the entire surface of the head to be depicted thus providing a
view of the voltage distribution (topographies) over the scalp for
each experimental condition as a function of time. Visual analysis
of the scalp maps provided four focuses of the EEG, one bilateral
including fourteen electrodes (showed in triangles in Figure 4.1B)
lasting from 180 to 220 ms, another posterior, including seven
parieto-occipital electrodes (shown in circles in Figure 1B), lasting
from 205 to 275 ms after target presentation, a third, more central
focus, including thirteen parietal electrodes (shown in squares in
Figure 3.1B) lasting from 360 to 400 ms after the presentation of
the target, and a frontal focus of twelve electrodes (showed in
rhombuses in Figure 1B) lasting from 260-300 ms .
UNCONSCIOUS CSPCE IN CONFLICT & ERPs
157
Figure 4.1. (A) Trial sequence of events. (B) Groups of electrodes selected for
N170 (triangles), P2 (circles), P3 (squares) and N2 (rhombuses) components.
Latencies and voltage analyses were performed on a 40 ms
window centered at the N170 (200 ms) of the grand-average
waveform for the bilateral location, on a 70 ms window centered at
the P2 peak (240 ms) for the posterior location, on a 40 ms window
centered at the P3 peak (380 ms) for the central location, and on a
40 ms window centered at the N2 peak (280 ms) for the frontal
location. For each participant, trials that did not meet the criteria
regarding the amount of artefacts and bad channels were removed.
Mean amplitude voltages averaged over the selected channels and
SE II:EXPERIMENTO 2
158
time windows were submitted to repeated-measures ANOVAs with
Task, between-subject,, and Incongruency Percentage, Previous trial
congruency and Congruency, within-subjects, as factors. The same
ANOVA was used with subjects’ accuracy and RT.
3. Results
3.1 Behavioral results
Overall accuracy was very high (99%) and no significant
differences in accuracy was observed between experimental
conditions and tasks. Only correct responses were considered in the
RT and ERPs analyses. RTs shorter than 300 ms or longer than
1000 ms and post-error trials were excluded (3%).
3.1.1. Threshold setting
The average presentation time of the masked words at
threshold value was 20.75 ms (SD= ±2.70 ms) in the non-emotional
group and 19.75 (SD= ±4.71 ms) in the emotional group. A
comparison between these two threshold values showed no
significant differences between groups t > .1
3.1.2. RT analysis.
A significant main effect of task was found in the RT
analysis F(1,22)= 6.7229, p < .05, showing larger RT for the
emotional task (Mean= 631 ms, SD= ±65.173) than for the non-
emotional task (Mean= 567 ms, SD= ±54.069). We also found a
significant interaction of Incongruency percentage x Current trial
congruency F(1,22)= 4.7949, p< .05. (see Fig. 4.2.)
UNCONSCIOUS CSPCE IN CONFLICT & ERPs
159
Figure. 4.2. Interaction effect Incongruency percentage x Current trial
congruency in the RT.
Planned comparisons showed a close to significance
difference between congruent and incongruent trials when
incongruency percentage was low (33%), F(1,22)= 4.2536, p= .051
with larger RT for incongruent trials (Mean= 602, SD= ±73.137)
than for congruent trials (Mean= 596, SD= ±64.984). On the other
hand no significant differences between congruent (Mean= 601,
SD= ±66.088) and incongruent trials (Mean= 597, SD= ±67.723)
were found when incongruency percentage is high (66%).
A significant second order interaction Incongruency
percentage x Previous trial congruency x Current trial congruency
F(1,22)= 4.4971, p< .05 showed the expected Gratton effect when
incongruency percentage was low (33%) (see Fig. 4.3.).
SE II:EXPERIMENTO 2
160
Figure 4.3. Interaction effect Incongruency percentage x Previous trial
congruency x Current trial congruency in the RT.
As planned comparisons showed, when incongruency
percentage was low (33%) a close to significant difference between
congruent and incongruent trials was found if they were preceded
by a congruent trial F(1,22)= 4.1902, p= .052, showing larger RT
for incongruent (Mean= 604, SD= ±73.127) than for congruent
trials (Mean= 568, SD= ±54.887). No significant differences
between congruent (Mean= 598, SD= ±66.227) and incongruent
(Mean= 600, SD= ±74.673) trials were found when preceded by an
incongruent trial.
Reversed pattern of results were found in a high
incongruency context (66%). In this case congruent trials preceded
by a congruent trial (Mean= 605, SD= ±66.298) showed larger RT
than incongruent trials preceded by congruent trials (Mean= 596,
SD= ±65.039) F(1,22)= 5.521, p< .05 but these differences
disappeared when congruent (Mean= 597, SD= ±67.008) and
UNCONSCIOUS CSPCE IN CONFLICT & ERPs
161
incongruent trials (Mean= 598, SD= ±71.685) were preceded by an
incongruent trial.
3.2.EEG analysis
3.2.1. Target analysis
3.2.1.1.N170 (bilateral group of electrodes)
Latency analysis
The analysis was conducted on a 40 ms window centered on
a negative peak at 200 ms. A close to significance interaction effect
was found between Incongruency percentage x Current trial
congruency F(1,22)=3,556, p= .072 (see Fig. 4.4).
Figure 4.4. Waves and graphic representation of the latency interaction effect
Incongruency percentage x Current trial congruency effect in N170 latency.
Planned comparisons showed only a significant difference
between congruent trials (Mean= 191.698, SD= ±11.192) and
incongruent trial (Mean= 189.411, SD= ±9.819) in a low
SE II:EXPERIMENTO 2
162
incongruency context (33%) F(1,22)= 6.49, p< .05. Likewise an
additional close to significance interaction between Previous trial
congruency and Current trial congruency appeared, F(1,22)= 3.295,
p= .083 (see Fig. 4.5)
Figure 4.5. Waves and graphic representation of the interaction effect Previous
Congruency x Congruency in N170 latency.
Only the difference between congruent (Mean= 191.438,
SD= ±10.852) and incongruent trials (Mean= 188.839, SD= ±8.971)
when previous trial was incongruent turned up significant F(1, 22)=
5.45, p<.05. No other significant difference was found. However,
this interaction was modulated by Incongruency percentage as
shown by a close to significant second order interaction, F(1,22)=
3.7714, p= .065. As can be seen in Figure 4.6, the difference
between congruent (Mean= 192.448, SD= ±11.425) and
incongruent trials (Mean= 187.563, SD= ±8.030) when previous
trial was incongruent was significant F(1,22)= 6.91, p< .05, when
the incongruency context was low (33%). Also in this low
UNCONSCIOUS CSPCE IN CONFLICT & ERPs
163
incongruency context we found that the incongruent following
another incongruent trial (Mean=187.563, SD= ±8.030) peacked
earlier than incongruents following a trial (Mean= 191,260, SD=
±11.110), F(1, 22)= 5.652, p< .05.
Figure 4.6. Graphic representation of the interaction effect Incongruency
percentage x Previous trial congruency x Current trial congruency for the N170
latency.
Finally an additionally significant second order interactions
was found between Task x Incongruency percentage x Current trial
congruency F(1,22)= 5.1354, p< .05 (see Fig.6). This interaction
was due to a significant F(1,22)= 4.64, p< .05 decrease in latency
for previous congruent trials when the incongruency context was
high (66%) (Mean= 188.135, SD= ±9.950) compared to when it
was low (33%) (Mean= 190.260, SD= ±9.856) only for the non-
emotional task (See Figure 4.7).
SE II:EXPERIMENTO 2
164
Figure 4.7. Graphic representation of the interaction effect Task x Incongruency
percentage x Previous trial congruency in the N170 latency.
Amplitude analysis
Voltage analysis showed an significant interaction between
Task and Hemisphere F(1,22)= 13.501, p< .01. In the right
hemisphere, the difference in amplitude between the non-emotional
(Mean= -1,152, SD= ±3,475) and the emotional task (Mean= 1,019,
SD= ±1.85) was close to significance, F(1,22) = 3,635, p = .07. But
there were no difference between non-emotional and emotional
tasks in the left hemisphere.
We also found a significant second order interaction effect,
hemisphere x Incongruency percentage x Current trial congruency
F(1,22)= 4.3916, p< .05. The only close to significance effect was
found in the left hemisphere and was due to a decrease in negativity
for incongruent trials in a high incongruency (66%) (Mean= -0,051,
SD= ± 2,130) compared with a low incongruency (33%) (Mean= -
0.232, SD= ± 2,245) context. No other effect reached significance.
UNCONSCIOUS CSPCE IN CONFLICT & ERPs
165
3.2.1.1. P2 (Posterior group of electrodes)
Inspection of the scalp topographies revealed a P2 positive
polarity (see Fig. 4.8) lasting from 205 to 275 ms after target
presentation over a set of parieto-occipital electrodes (circles in Fig.
4.1B).
Figure 4.8. Topographies corresponding to P2 peak (240 ms)
Latency analysis
A significant interaction between Task and Incongruency
percentage appeared, F (1,22) = 4,74, p < 0.5. Only the difference
in the emotional task between the 33% (Mean = 223,39 , SD = 8,09)
and 66% (Mean = 220,06 , SD = 9,87) conditions approached
SE II:EXPERIMENTO 2
166
significance, F(1,22) = 4,21, p = 0.052. A three-way interaction
Task x Previous trial congruency x Current trial congruency was
also significant, F (1,22) = 4,74, p < 0.05. Planned comparisons
showed that only in the emotional task there was a marginally
significant increase in latency (F1,22) = 3.564, p= 0.07 for the
incongruent trials following another incongruent trial (Mean =
222,64 SD = 10,42) compared to incongruent trials following a
congruent trial (Mean = 220,33, SD = 8,15).
Amplitude analysis
Waveforms analysis sowed peak amplitudes around 240 ms,
and amplitudes were analyzed using adaptive means in windows of
70 ms around the appropriate peak. The amplitude in the 66%
incongruent trial condition (mean 66% = 3,998, SD = ± 3,277) was
higher than the amplitude in the 33% incongruent trial condition
(mean 33% = 3,803, SD = ± 3,291). This difference was found
significant, F(1,22) = 4.688, p < .05.
A significant second order interaction Task x Previous trial
congruency x Congruency was also found F(1,22)= 6.3213, p< .05.
(see Fig 4.9.).
UNCONSCIOUS CSPCE IN CONFLICT & ERPs
167
Figure 4.9. Waves and graphic representations of the interaction effect Task x
Previous trial congruency x Current trial congruency for the P2 amplitude.
SE II:EXPERIMENTO 2
168
As can be seen, in the non-emotional task, though it was
only marginally significant, incongruent trials (Mean= 3,982, SD=
± 3,721) showed less amplitude than congruent trials (Mean= 4,305,
SD= ± 3,757) when they were preceded by an incongruent trial.
F(1,22)= 3.508, p= .074. On the other hand, in the emotional task
this pattern of results is produced when previous trial is congruent
(Mean= 3,603, SD= ± 2,949) (Mean= 4,059, SD= ± 2,805) F(1,22)=
4.644, p< .05.
3.2.1.3. N2 (frontal group of electrodes)
Inspection of the scalp topographies showed a negative
polarity over a frontal set of electrodes (rhombuses in Fig. 3.1B) in
a window from 260 ms to 300 ms centered at 280 ms over a set of
frontal (see Figure 4.10).
UNCONSCIOUS CSPCE IN CONFLICT & ERPs
169
Figure 4.10.: Topographies corresponding to the N2 component.
Latency analysis
No main or interaction effects were found.
Amplitudes analysis
No significant main effects were found. But a significant
interaction effect between Previous trial congruency and Current
trial congruency F(1,22)= 4.7570, p< .05 (see Fig. 4.11) turned up.
SE II:EXPERIMENTO 2
170
Figure 4.11. Waves and graphic representations of the Interaction effect
Previous trial congruency x Current trial congruency for the N2 amplitude.
Planned comparisons showed a significant difference in
amplitude for congruent (Mean= -3.448, SD= ± 2.075) and
incongruent trials (Mean= -3.097, SD= ± 2.152) when previous trial
was incongruent F(1,22)= 5.1552, p< .05 compared with when
previous trial was congruent (Mean= -3.237, SD= ± 2.118, for
incongruent), (Mean= -3.378, SD= ± 2.210, for congruent). In
addition, incongruent trials showed a close to significance less
UNCONSCIOUS CSPCE IN CONFLICT & ERPs
171
negative when previous trial was incongruent than when previous
trial was congruent F(1,22)= 4.0597, p= .057.
3.2.1.4. P3 (central group of electrodes)
Inspection of the scalp topographies showed a positive
polarity over a centro-parietal set of electrodes (squares in Fig. 1.B)
in a window from 360 ms to 400 ms centered at 380 ms (see Fig
4.12).
Figure 4.12. Topographies corresponding to P3 peak (380 ms)
Latency analysis
No significant main or interaction effects were found.
SE II:EXPERIMENTO 2
172
Amplitude analysis
A significant main effect of Current trial congruency was
found for the P3 component F(1,22)= 4.5237, p< .05 with less
positivity for incongruent trials (Mean= 4,749, SD= ± 2.180)
compared with congruent trials (Mean= 4,903, SD= ± 2.187). We
also found a close to significance interaction effect between
Previous trial congruency and Current trial congruency F(1,22)=
3.6828, p= .068 (See Fig. 4.13).
Figure 4.13. Waves and Graphic representations of the close to significance
interaction effect Previous trial congruency x Current trial congruency for the P3
amplitude.
UNCONSCIOUS CSPCE IN CONFLICT & ERPs
173
Planned comparisons showed significant differences
between congruent (Mean= 4.990, SD= ± 2.202) and incongruent
trials (Mean= 4.627, SD= ± 2.203) when they were preceded by
another incongruent trial F(1,22)= 6.1521, p< .05. No significant
differences were found when previous trial was congruent. We also
found a significant third order interaction Task x Incongruency
percentage x Previous trial congruency x Current trial congruency
effect F(1,22)= 4.848, p< .05. To analyze this interaction effect we
took emotional and non-emotional tasks separately. Analyzing the
emotional task we found a close to significance second order
interaction Incongruency percentage x Previous trial congruency
Current trial congruency effect F(1,11)= 4.2901, p= .062 (see Fig.
4.14 y 4.15).
Planned comparisons showed a significant difference
between congruent (Mean= 5,061, SD= ± 2.419) and incongruent
trial (Mean= 4.546, SD= ± 2.527) when previous trial was
incongruent, in the high incongruency context (66%), F(1,11)=
8.766, p< .05. In addition, the difference in amplitude between
incongruent following congruent (Mean= 5,089, SD= ± 2.431) and
incongruent following incongruent trials was also significant
F(1,11)= 6.4295, p< .05. However no significant difference
appeared in low incongruency context (33%)
SE II:EXPERIMENTO 2
174
Figure 4.14. Waves of the second order interaction effect Incongruency
percentage x Previous trial congruency x Current trial congruency in emotional
task for P3 amplitude.
UNCONSCIOUS CSPCE IN CONFLICT & ERPs
175
Figure 4.15. Graphic representation of the second order interaction effect
Incongruency percentage x Previous trial congruency x Current trial congruency
in emotional task for P3 amplitude.
On the other hand, in the non-emotional task, we found a
close to significance interaction effect Previous trial congruency
and Current trial congruency F(1,22)= 3.9167, p= .073 (See Fig.
4.16)
Figure 4.16. Interaction effect Previous trial congruency x Current trial
congruency in P3 amplitudes for the Non-Emotional task for P3 amplitude.
Planned comparisons showed significant differences
between congruent (Mean = 4,90, SD = 2,50) and incongruent trials
SE II:EXPERIMENTO 2
176
(Mean = 4,65, SD = 1,92) only when previous trial was incongruent
F(1,11)= 6.5858, p<.05.
4. Discussion
The main purpose of this experiment was to study the
cognitive and emotional conflict effects under unconscious
conditions and to observe how the proportion of incongruency
modulates these effects when it is associated to a particular context.
Thus, we tried to replicate the results of Panadero et al. (2015),
concerning the CSPCE, in a different Stroop-like cognitive task in
which participants had to identify the gender of male or females
faces. At the same time, we were interested in testing the
similarities and/or differences of a hypothetical CSPCE related to
emotional control, using the same Stroop-like face categorization
task but, in this case, asking the participants to identify the emotion
expressed either happy or frighten faces. In addition, we also
analyzed the possible effect due to the congruent or incongruent
character of the previous trials, in an attempt to shed some light on
the likely relationship between this variable and proportion of
incongruency. To insure the unconscious nature of Stroop conflict,
the labels indicating either the gender or the emotion of faces were
presented before the faces and double masked by a preceding and
following set of characters. This set, in turn, was associated to and
different for each proportion of incongruency.
We employed a subjective-threshold criterion of awareness
determined by word presentation time. Below this threshold
participants reported to be unable to perceive the words. In fact, the
UNCONSCIOUS CSPCE IN CONFLICT & ERPs
177
mean presentation time was 20.75 ms in the non-emotional group
and 19.75 in the emotional group, with no significant difference in
threshold value between the two. These presentation times are
similar than those used in other masking experiments (Greenwald,
Klinger, & Liu, 1989; Klapp, 2007; Panadero et al., 2015; Ortells,
Marí-Beffa & Plaza-Ayllón, 2012). Additionally, when subjects
were asked after the experiment if they had been able to process any
word, the general answer was negative and none could establish any
relation between the masks and the proportion of incongruency.
Even more, all subjects reported equal difficulty across blocks
independently of the mask type. Hence, it seems safe to conclude
that, under subjective threshold conditions, our participants were
unaware of the prime words and of the relationship between masks
and proportion of incongruency.
Our RT data showed that the emotional task was
significantly slower than the non-emotional task. These results
agree with Egner et al. (2008) and Panadero & Tudela (to be
submitted), who used a face categorization task and reported the
same difference as well. However Alguacil et al. (2013) and
Oschner et al. (2008), using words in an Eriksen flankers task, did
not find difference in RT between cognitive and emotional tasks.
Therefore it is likely that these discrepant results may be due to the
type of stimuli on which the response is made. It appears that
identifying a male or a female face is relatively easy compared with
discriminating the emotional expression of those faces; while
identifying the face gender may be based on global characteristics
of the face, discriminating its emotional expression may require to
SE II:EXPERIMENTO 2
178
focus attention on specific facial features like eyes, nose and mouth.
In fact, influential models of face perception (Bruce & Young,
1986; Haxby & Gobini, 2011) consider gender to be part of a face
identity (core system according to Haxby & Gobini, 2011) because
it is a permanent characteristic and hence is processed at early
stages of face perception. However emotional facial expression,
being a changeable characteristic, requires additional information
and is processed at later stages of face perception (extended system
for Haxby and Gobini, 2011).
We also found the significant interaction between
Incongruency percentage and Congruency characteristic of conflict
adaptation; in a context of low incongruency (33%), incongruent
were slower than congruent trials but, in a context of high
incongruency (66%), incongruent tended to be faster than congruent
trials. These results replicate those obtained by Panadero et al.
(2015), who used a color-word-color patch cognitive Stroop task,
and extends their results to emotional conflict as well, since the
significant interaction in our data did not appear modulated by the
type of task. Therefore our results may be considered an additional
demonstration of unconscious cognitive CSPCE, and, to our
knowledge, the first demonstration of unconscious emotional
CSPCE. Previous research (Crump, Vaquero & Milliken, 2008;
Blais, Harris, Guerrero, & Bunge, 2012) had shown that CSPCE
might be independent of subjects’ awareness of the incongruency
proportion in situations where no masking procedure had been used.
In the present research, by means of masking, subjects were made
unaware of the congruency relation between the masked word and
UNCONSCIOUS CSPCE IN CONFLICT & ERPs
179
either the gender or the emotion of faces. In addition within each
block of trials the percentage of congruent and incongruent trials
was the same. Under these conditions, awareness of the proportion
of incongruency associated with each mask seems very unlikely.
In the present experiment, Previous trial congruency
produced an effect that was modulated by Incongruency percentage.
The Stroop effect decreased in size from the condition where the
previous trial was congruent to the condition where the previous
trial was incongruent, but this decrement occurred only in the low
incongruency context (33%). In the high incongruency context
(66%), when the previous trial was congruent a reversed Stroop
effect appeared, where congruent were slower than incongruent
trials; on the other hand, when the previous trial was incongruent,
no difference between congruent and incongruent trials turned up.
If we assume that the manipulation of both variables, Previous trial
congruency and Incongruency percentage, influence the Stroop
effect in the same way, both tend to decreasing it (Botvinick et al.,
2001), it is reasonable to suppose that the effect of one variable, say
Previous trial congruency, may more clearly show up in a context
where the effect of the other variable, say Incongruency percentage,
is relatively weak, as it is the case in the 33% incongruency
condition. However in a context where the Stroop effect is already
reduced, as it is the case in he 66% incongruency context, the effect
of Previous trial congruency in bound to be faint, because there is
little room for the Stroop effect to be further reduced. In addition,
the inverted Stroop effect, found when the previous trial was
congruent in the 66% condition, may easily be due to the unlikely
SE II:EXPERIMENTO 2
180
presentation of two consecutive congruent trials in a context where
the dominant expectation favors an incongruent trial. It is
interesting to note that the above mentioned effects were
independent of the type of task; no difference was modulated by the
conflict cognitive or emotional nature. Nevertheless, our second
order significant interaction between Previous trial congruency,
Incongruency percentage and Current Congruency is an intriguing
result in need of further replication.
With respect to our electrophysiological results, we found
significant effects in both, latency and amplitude, of the N170
component. The N170 latencies showed close to significance
interactions of Congruency with Incongruency percentage as well
as with Previous trial congruency. However these effects were
clarified by a second order interaction among the three variables.
As can be seen in Figure 6, only in the low incongruency context
(33%) appeared a significant decrease in N170 latency for
incongruent trials when they were preceded by an incongruent
compared with a preceding congruent trial. This Gratton effect
resembles that found in RT and indicates again that the effect of a
previous congruent or incongruent trial depends upon the
incongruency context; in our experiment, when the incongruency
context was high (66%) no significant effect appeared. Similar to
our RT results, these effects were independent of task type.
However the type of task did showed up interacting with Previous
trial congruency and Incongruency percentage; only for the non-
emotional task, previous congruent trials produced shorter latencies
in the high (66%) than in the low (33%) incongruency context.
UNCONSCIOUS CSPCE IN CONFLICT & ERPs
181
Nontheless this interaction provides little information about the
main purpose of the present experiment since the Stroop effect was
not involved in it.
Amplitude analysis for the N170 mainly showed
lateralization effects. On the one hand, only in the right hemisphere
the non-emotional tended to show larger negativity than the
emotional task. On the other hand, only in the left hemisphere there
were indications of conflict adaptation since the negativity of the
incongruent trials tended to change as a function of the type of
previous trial, this negativity was shorter after an incongruent than
after a congruent trial. This result agrees, but only in part, with that
reported by Alguacil et al. (2013). They found conflict adaptation
on the N170 independent of hemisphere and with voltage values
more negative for incongruent trials following identical trials than
for those following a congruent trial. In our experiment, conflict
adaptation was lateralized and the direction of change for
incongruent trials was opposite to that found by Alguacil et al. It
seems that the direction of change induced by the type of previous
trial on incongruent trials negativity is not clear. We will take over
this issue after analyzing the remaining ERP components.
In sum, the analysis of the N170 component provided
interesting information with respect to the aims of the present
research. As in the RT analysis, unconscious conflict adaptation
was found in the N170 latencies but depending upon the
incongruency context. It was present in a context of mostly
congruent trials but not in a context of mostly incongruent trials. In
SE II:EXPERIMENTO 2
182
turn, N170 amplitude analysis showed conflict adaptation only in
the left hemisphere but the negativity of incongruent trials
decreased, rather than increased, after an identical trial compared
with after a congruent one.
In the latencies of the P2 component, conflict adaptation
appeared close to significance only for the emotional task but
latency for incongruent trials tended to increase instead of decrease
when they were preceded by an identical trial compared to when
they were preceded by a congruent trial. The amplitude analysis
replicated previous results by Panadero et al. (2015) showing a
significant main effect of Incongruency percentage; the amplitude
of the mostly incongruent trial condition was higher than the
amplitude of the mostly congruent trial condition. In addition, a
different effect of Previous trial congruency showed up depending
upon type of task. In the non-emotional task, a close to significance
reversed Stroop effect appeared under the previous incongruent
trial condition while in the emotional task the reversed Stroop effect
was significant and appeared under the previous congruent trial
condition.. In sum, the only consistent result found in the P2
component was the main effect of Incongruency proportion
replicating previous results by Panadero et al. (2015). The
remaining results look rather unstable; though the type of task
seems to modulate conflict adaptation, the way it appears to do it
does not follow any known or understandable pattern of data.
In contrast with the just mentioned results, our data
concerning N2 showed a clear unconscious conflict adaptation
UNCONSCIOUS CSPCE IN CONFLICT & ERPs
183
effect that was independent of both, type of task and percentage of
incongruency. Interestingly, the adaptation effect showed a
decrease in negativity for incongruent trials from the previous
congruent to the previous incongruent trial condition. This result
agrees with Clayson and Larson (2011) who, using a modified
Eriksen flanker task in which participants had to identify the
direction of a central arrow pointing to the right or to the left,
reported a decrease of the N2 amplitude for incongruent trials
following another incongruent trial compared to those following a
congruent one. Likewise Panadero and Tudela (to be submitted), in
an experiment using the same tasks than in the present experiment,
found less negativity for incongruent than for congruent trials in the
N2 amplitude but only in the emotional task, in the cognitive task
the effect was reversed and the negativity was larger for
incongruent than for congruent trials. Consequently, whether
conflict adaptation should show up as a decrease in negativity for
incongruent trials as a function of the type of the previous trial, or
not, may depend on variables such as type of task. To what extent
this line of reasoning may also be extended to our results in the
N170 amplitude will need further investigation.
The P3 amplitude showed a significant third order
interaction with different patterns of results for the emotional and
non-emotional tasks. In the emotional task there was a significant
Gratton effect but only in the context of high incongruency (66%).
We found no effect in the low incongruency (33) context. In the
non emotional task, the Gratton effect was significant and
independent of incongruency context. The results for the
SE II:EXPERIMENTO 2
184
emotional task show a modulation of the Gratton effect by the type
of context opposite to the way in which RT and the N170 latencies
were modulated by the same variable. In the RT and N170
latencies, the Gratton effect appeared in a context of low
incongruency (33%) while in the P3 it appeared in a context of high
congruency (66%). Though this difference is not easy to explain,
may be related to the characteristics of the P3 amplitude providing
room for greater differences between congruent and incongruent
trials than it was the case for RT. Thus, the Stroop effect would
have a greater range for change in the P3 amplitude than in RT. On
the other hand, it is well known that the P3 amplitude is inversely
related to the probability of task-defined stimulus class (Kutas,
McCarthy, & Donchin, 1977). Consequently, in our experiment,
the greater the probability of an incongruent trial, the shorter the P3
amplitude for these trials would be. Thus, the 66% percentage of
incongruency context would tend to decrease the P3 amplitude of
incongruent and increase the amplitude of congruent trials. In
addition, a previous incongruent trial would tend to further decrease
the amplitude of incongruent trials, resulting in a significant Gratton
effect when the incongruency context is high (66%). On the
contrary, the 33% incongruency context would tend to decrease the
amplitude of congruent and increase the amplitude of incongruent
trials, while the effect of a previous incongruent trial would tend to
decrease the amplitude of the incongruent trials. In this context the
P3 amplitude of incongruent trials would undergo the influence of
two tendencies that are opposite to each other. As a result,
predictions about what to expect in P3 amplitude in the 33%
UNCONSCIOUS CSPCE IN CONFLICT & ERPs
185
incongruency context are almost impossible. The fact that the
above-discussed two-way interaction only appears in the emotional
task, may be related to the greater difficulty of this task, as indicated
by our RT data. As P3 amplitude is known to increase with task
difficulty (Isreal, Chesney, Wickens, & Donchin, 1980), it may be
the case that only the emotional task provided the appropriate
amplitude background for the complexities of the two-way
interaction to show up. In any case, the significant three-way
interaction found in the present experiment under masking
conditions, is in itself a valuable empirical finding that deserves
further investigation.
In summary, the present experiment clearly demonstrated
unconscious conflict adaptation in the RT measures, as well as in
several ERP components like N170 latencies, and N2 and P3
amplitudes. These results replicate those reported by Panadero et
al. (2015), though in the Panadero et al. experiment the type of
conflict investigated was only cognitive and conflict adaptation was
induced by the manipulation of context incongruency but not by
previous congruency. In this experiment context incongruency
interacted with the type (congruent or incongruent) of previous trial
and also with the type (cognitive or emotional) of task. So, a
contribution of this experiment that deserves to be pointed out is the
wide interplay of variables able to modulate the Stroop effect and
the conflict adaptation effect. Finally, it should not be overlooked
that all the effects reported in this experiment took place in a
situation in which participants were unaware of the labels
associated with the faces and of the relations between masks and
SE II:EXPERIMENTO 2
186
incongruency percentage. We may conclude that in this experiment
unconscious conflict adaptation has been demonstrated for both,
cognitive and emotional, types of conflict.
Capítulo 5
DISCUSIÓN GENERAL
En términos generales, el objetivo principal de esta tesis ha
sido estudiar los procesos de adaptación al conflicto: el efecto de
proporción de congruencia asociada a un determinado contexto
(CSPCE) y el efecto asociado a la secuencia de ensayos, conocido
como efecto Gratton. Hemos estudiado estos efectos en situaciones
fundamentalmente de no consciencia, observando el curso de la
actividad neural asociada a dichos procesos. Además hemos querido
estudiar las semejanzas o diferencias de las estrategias de resolución
de conflicto, tanto cognitivo como emocional, mediante el análisis
del registro de la actividad eléctrica del cerebro.
Con este fin, se diseñaron dos series experimentales. La
primera serie constaba de un solo experimento cuyo propósito
principal fue estudiar los procesos de control cognitivo a nivel no
consciente, así como ver la modulación que sufrían dichos procesos
de control al manipular la proporción de incongruencia, cuando ésta
se asociaba a un determinado contexto. Para ello se utilizó una
variante del paradigma clásico de Stroop (Stroop, 1935), en el cual
se enmascaraba la palabra de color usando dos tipos de máscaras
que se asociaban sistemáticamente con una proporción de
incongruencia alta (66%) o una proporción de incongruencia baja
(33%). Procuramos mantener activa la expectativa de presentación
de dicha palabra de color, mediante la calibración del tiempo de
presentación de dicha palabra para cada participante. Durante todo
el procedimiento se registró la actividad eléctrica cerebral usando
técnicas electrofisiológicas de alta densidad, concretamente
potenciales corticales evocados (ERP). Los resultados del primer
experimento mostraron evidencia a favor del carácter no consciente
CAPÍTULO 5
190
de los procesos de control cognitivo asociados a la tarea Stroop, y
de la existencia de CSPCE a nivel no consciente, como se
desprende del análisis del tiempo de reacción (RT) y de diferentes
componentes de la actividad eléctrica del cerebro (ERPs). Así a
nivel de RT se observó una disminución del efecto Stroop en
aquellas situaciones en las que la proporción de incongruencia era
más alta (66%), lo cual es un claro efecto de adaptación al conflicto
(Logan, et al., 1984; Jacoby et al., 2003; Crump, Gong y Milliken,
2006) que, en este caso, se produjo en condiciones de no
consciencia de la palabra previa que designaba al color. Este efecto
se reflejó a su vez en los potenciales corticales N2 y P3 cuya
amplitud varió también en función de la proporción de
incongruencia. A su vez, el análisis del componente P2 mostró
evidencia a favor de la existencia de un proceso de aprendizaje
implícito de la asociación de las máscaras con sus respectivos
porcentajes de incongruencia a lo largo del experimento
En la segunda serie experimental, compuesta por dos
experimentos, ampliamos nuestros objetivos para estudiar además
del control cognitivo, el control emocional. En el primer
experimento se estudió el control consciente mientras que en el
segundo experimento estudiamos el control inconsciente. El
paradigma común a los dos experimentos de esta serie fue una
variante de la tarea de categorización de caras (Egner, Etkin, Gale,
& Hirsch, 2008; Etkin, Egner, Peraza, Kandel, & Hirsch, 2006), en
la cual los participantes debían clasificar una serie de caras en
función del género de estas (en la versión cognitiva) o bien en
función de la emoción que expresaban (en la versión emocional). El
DISCUSIÓN GENERAL
191
objeto de elegir este paradigma fue igualar la tarea para el estudio
de ambos tipos de conflicto al mismo tiempo que diferenciar las
respuestas sobre el estímulo objetivo que en un caso eran de
carácter cognitivo y en el otro emocional. En el primer experimento
de la serie el efecto de adaptación al conflicto se estudió mediante el
análisis de la influencia de la congruencia previa pero no se
manipuló la proporción de incongruencia. De esta forma quisimos
estudiar, utilizando potenciales corticales, la situación analizada
mediante fMRI por Egner et al. (2008) y compararlos con los
resultados de Alguacil et al. (2013) que, utilizando potenciales
corticales, habían estudiado conjuntamente el control cognitivo y
emocional empleando una tarea de flancos de Eriksen. Los
resultados comportamentales replicaron los obtenidos
anteriormente, tanto para el conflicto cognitivo (McCleod, 1991),
como para el conflicto emocional (Mckenna& Sharma, 1995; Etkin,
Egner, Peraza, Kandel& Hirsch, 2006; Bush., Whalen., Rosen.,
Jenike., McInerney., & Rauch., 1998; Hayward., Goodwin., &
Harmer., 2004), como para ambos conjuntamente (Alguacil, et al.,
2013; Egner et al., 2008). Con respecto a los potenciales corticales
analizados, encontramos un efecto de adaptación al conflicto en N2
y P3, mostrando el N2 una disociación entre conflicto cognitivo y
emocional que comentaremos más adelante.
En el segundo experimento de la segunda serie utilizamos la
misma tarea que en el experimento anterior pero introdujimos dos
cambios fundamentales. El primero consistió en estudiar la
adaptación al conflicto manipulando también el porcentaje de
incongruencia, cuyo efecto habíamos demostrado en el experimento
CAPÍTULO 5
192
de la primera serie. Como en este experimento, asociamos el
porcentaje de incongruencia al tipo de máscara que se utilizó para
impedir el procesamiento consciente de las palabras que etiquetaban
el género o la emoción de las caras. De esta forma el segundo
cambio introducido respecto al experimento anterior en la serie fue
el enmascaramiento de dichas palabras. Los resultados
comportamentales mostraron de nuevo un CSPCE inconsciente,
independientemente del tipo de tarea. También mostraron una
interacción entre porcentaje de incongruencia, congruencia previa y
efecto Stroop, de forma que solamente en el contexto de baja
incongruencia se mostró el efecto Gratton. Con respecto a los
potenciales corticales analizados, efectos de adaptación al conflicto
aparecieron no sólo en las latencias sino también en las amplitudes
tanto del N170 como del N2 y P3. El comentario de estos efectos
se retomará más adelante. El P2 replicó el resultado del primer
experimento mostrando una sensibilidad diferencial al porcentaje de
incongruencia.
Las principales aportaciones de esta tesis están relacionadas
fundamentalmente con dos aspectos principales: los efectos de
adaptación al conflicto y el procesamiento no consciente de los
procesos relacionados con esos efectos.
El efecto de adaptación al conflicto
Con respecto a los efectos de adaptación al conflicto, los
resultados han puesto de manifiesto que tanto la manipulación del
porcentaje de incongruencia como la manipulación de la
congruencia del ensayo previo son variables que modulan el efecto
DISCUSIÓN GENERAL
193
Stroop, reflejando esta modulación tanto en las medidas
comportamentales como en las medidas relacionadas con algunos
componentes de los potenciales corticales evocados.
Los resultados comportamentales han mostrado un efecto
significativo de adaptación al conflicto en dos de los tres
experimentos de la tesis. En la primera serie experimental el efecto
fue producido por la manipulación de la proporción de
incongruencia y en el tercero fue producido por la variable anterior
y además por la congruencia previa. Curiosamente en los dos
experimentos en los que se ha encontrado adaptación al conflicto, la
presentación de las palabras estímulo-previo estaban enmascaradas.
Dicho con otras palabras el efecto de adaptación al conflicto ha
aparecido en situaciones de procesamiento no consciente de los
estímulos previos. Dentro de las limitaciones de nuestros
conocimientos, esta es la primera vez que se demuestra un CSPCE
no consciente, cognitivo en el primer experimento y tanto cognitivo
como emocional en el tercer experimento. A nivel consciente estos
efectos han sido ampliamente establecidos para el conflicto
cognitivo (Crump, Gong, y Milliken, 2006; Crump, Vaquero &
Milliken, 2008; Jacoby, Lindsay, and Hessels, 2003; Logan,
Zbrodoff, & Williamson, 1984), y también, aunque en menor
medida para el conflicto emocional (Alguacil et al.; 2013; Egner et
al., 2008, Zhu, Zhang, Wu, Luo y Luo, 2010). En nuestro segundo
experimento, no obtuvimos dicho efecto ni cognitivo ni emocional,
aunque los potenciales corticales evidenciaron la presencia de un
efecto de adaptación al conflicto. Nuestro tercer experimento
mostró una interacción de segundo orden que reflejaba la
CAPÍTULO 5
194
modulación conjunta del efecto Stroop por el porcentaje de
incongruencia y por la congruencia previa, de modo que el efecto
Gratton se localizó en el contexto de baja incongruencia pero no en
el de alta incongruencia. Si suponemos, como suele ocurrir, que las
dos variables manipuladas tienden a afectar al TR de la misma
forma, este resultado puede deberse a que la probabilidad de que
aparezca el efecto Gratton sea mayor en un contexto en el que el
efecto del porcentaje de ensayos incongruentes sea menor, como se
ha explicado en la discusión del tercer experimento.
En relación con los potenciales corticales analizados,
nuestros resultados presentan una consistencia sistemática en los
tres experimentos, mostrando efecto de adaptación al conflicto tanto
en N2 como en P3. El N2 presentó una topografía frontal en los
experimentos primero y tercero, y una topografía parietal posterior
en el segundo. La adaptación al conflicto cognitivo presentó una
disminución en la negatividad de los ensayos incongruentes
respecto a la de los congruentes en el primer experimento y un
aumento de su negatividad en el segundo; en el tercer experimento,
independientemente del tipo de conflicto, la negatividad del N2 se
redujo en la misma dirección que en el experimento primero. En
cuanto al conflicto emocional, sólo en el segundo experimento la
negatividad de los ensayos incongruentes presentó una disminución
con respecto a la de los congruentes. Cabe la posibilidad de que la
direccional del cambio inducido en los ensayos incongruentes como
resultado de la adaptación al conflicto dependa del tipo de conflicto.
Así parecen sugerirlo los resultados de nuestro segundo
experimento, sin embargo, Clayson y Larson (2011) encontraron un
DISCUSIÓN GENERAL
195
decremento de la negatividad en una situación de conflicto
cognitivo. En consecuencia, parece que el problema de la
direccionalidad del cambio que cabe esperar en el N2 como
resultado de la adaptación al conflicto, necesita mayor
investigación.
En el componente P3 la adaptación al conflicto, tanto
cognitivo como emocional, siempre se manifestó como una
disminución en la positividad de los ensayos incongruentes con
respecto a los congruente. En el primer experimento, que sólo
estudió el conflicto cognitivo, se encontró un efecto significativo
de adaptación al conflicto. En el segundo experimento, el efecto
de adaptación al conflicto solamente apareció significativo en la
tarea non-emocional. Finalmente en el tercero, presentó un patrón
de resultados diferente dependiendo del tipo de tarea. En la no
emocional se encontró un efecto de la congruencia previa cercano a
la significatividad que era independiente del porcentaje de
incongruencia. En la tarea emocional, se encontró un efecto
Gratton significativo, pero únicamente cuando el contexto era de
alta incongruencia. Curiosamente este patrón, que implicaba al
porcentaje de incongruencia, a la congruencia previa y al efecto
Stroop, era diferente al presentado por la interacción de estas
mismas variables en el tiempo de reacción y en las latencias del
N170 del tercer experimento. En estos dos últimos casos el efecto
Gratton apareció cuando el contexto era de baja incongruencia y fue
independiente del tipo de tarea. No resulta fácil explicar esta
discrepancia de resultados. El razonamiento avanzado
anteriormente respecto al efecto encontrado en el tiempo de
CAPÍTULO 5
196
reacción se apoya principalmente en el limitado rango de variación
del efecto Stroop que parece razonable esperar cuando se mide en
tiempo de reacción. En el caso del P3, es más difícil establecer el
rango de variación del efecto Stroop aunque podemos razonar de la
siguiente manera.
Por una parte a nivel teórico, este componente ha sido
relacionado con la “adaptación al contexto” (Donchin & Coles,
1988) y con la asignación de recursos cognitivos a una tarea (Isreal,
Chesney, Wickens, & Donchin, 1980), de forma que la amplitud del
P3 aumenta cuando la tarea es difícil y necesita mayor asignación
de recursos. Por otra parte a nivel empírico, sabemos que la
amplitud del P3 es sensible a la probabilidad del objetivo (Duncan-
Johnson y Donchin (1977), sobre todo a la probabilidad de la clase
de estímulos definidos por la tarea (Kutas, McCarthy, & Donchin,
1977), de forma que su amplitud es inversamente proporcional a
dicha probabilidad.
En nuestro experimento podemos asumir que la tarea
emocional es más difícil que la no emocional y por tanto necesita
mayor cantidad de recursos. Este supuesto está fundamentado en el
hecho de que tanto en el segundo como en el tercer experimento
hemos obtenido un efecto principal de tarea en el TR, de forma que
la tarea emocional ha sido mas lenta que la no emocional. Podemos
por tanto asumir que la amplitud del P3 va a poder aumentar (Isreal,
Chesney, Wickens, & Donchin, 1980) y, en consecuencia, tener
mayor rango de variación en la tarea emocional que en la no
emocional. Por otra parte, el contexto de alta incongruencia (66%)
DISCUSIÓN GENERAL
197
introduce una alta probabilidad de ensayos incongruentes que va a
disminuir la amplitud de los ensayos incongruentes y a aumentar la
de los congruentes (Kutas, McCarthy, & Donchin, 1977)
permitiendo que los ensayos previos incongruentes produzcan su
efecto de adaptación al conflicto, disminuyendo aún más la
amplitud del P3 de los ensayos incongruentes. Sin embargo, el
contexto de baja incongruencia (33%) introduce una alta
probabilidad de ensayos congruentes que va a disminuir la amplitud
del P3 de estos ensayos y a aumentar la de los ensayos
incongruentes. En este caso, el efecto de un ensayo previo
incongruente influirá sobre la amplitud del P3 de estos ensayos en
sentido opuesto a la influencia del contexto.
Dicho con otras palabras, la tarea emocional exige más
recursos que la no emocional aumentando la amplitud del P3 de esta
tarea y proporcionando un margen de reducción de su amplitud a la
influencia de las otras dos variables. La aparición de la máscara
asociada al alto porcentaje de incongruencia introduce un contexto
de alta probabilidad de ensayos incongruentes que va a disminuir la
amplitud del P3 de esos ensayos por ser los más probables y a
aumentar la amplitud del P3 de los ensayos congruentes por ser los
menos probables. Finalmente, la expectativa inmediata provocada
por un ensayo previo incongruente disminuirá aún más la amplitud
del P3 para los ensayos incongruentes. Como consecuencia de la
interacción de estos factores, el efecto Gratton aparece significativo
en el contexto de incongruencia alta. En un contexto de
incongruencia baja (33%), la probabilidad más alta está asociada
con los ensayos congruentes que influirá reduciendo la amplitud del
CAPÍTULO 5
198
P3 asociado con estos ensayos y aumentando la amplitud del P3 de
los ensayos incongruentes. En esta situación la influencia de un
ensayo previo incongruente actuará sobre los ensayos incongruentes
disminuyendo la amplitud de su P3, es decir, en sentido opuesto a la
influencia del contexto. En estas condiciones resulta muy difícil
predecir las amplitudes finales del P3 de los ensayos congruentes e
incongruentes a la hora de poder producir un efecto Stroop. De
hecho en nuestro experimento no encontramos ningún resultado
significativo en el contexto de baja incongruencia.
Aunque este razonamiento no carece de fundamento, puede
resultar bastante especulativo, por lo que merece la pena resaltar el
valor que tiene en sí mismo el resultado empírico que hemos
obtenido en el tercer experimento, único en el que se ha podido
analizar conjuntamente el efecto de la proporción de incongruencia
y de la congruencia previa sobre el efecto Stroop. En tres
ocasiones, en el TR, latencias del N170 y amplitud del P3 hemos
encontrado significativa la interacción de segundo orden entre
porcentaje de incongruencia, congruencia previa y la congruencia
del ensayo actual, aunque con patrones de resultados diferentes.
Creemos que esta interacción de segundo orden, así como la de
tercer orden, que implicó también al tipo de tarea y que apareció en
la amplitud del P3, constituyen una aportación novedosa de nuestra
investigación que obliga a repensar la forma en que estas variables
modulan el efecto Stroop. La aportación tiene aún más valor si
tomamos en consideración que estos resultados han aparecido en
una situación experimental en la que los estímulos previos habían
sido enmascarados.
DISCUSIÓN GENERAL
199
Por lo que respecta a los resultados encontrados en los
potenciales corticales más tempranos, la información que nos han
proporcionado ha sido considerablemente más escasa. En el
segundo experimento, no se encontró efecto alguno en las latencias
del P1 ni del N170, y el análisis de las amplitudes sólo puso de
relieve una aproximación a la significatividad del efecto Gratton en
el P1. A su vez, en el tercer experimento solamente las latencias del
N170 mostraron la interacción de segundo orden a la que hemos
aludido anteriormente y cuyo patrón de resultados era similar al
encontrado en el TR, es decir el efecto Gratton se encontró en el
contexto de baja frecuencia (33%) pero no en el de alta frecuencia
(66%). Dado que el efecto se ha encontrado en las latencias, hemos
preferido explicar esta interacción de segundo orden en los mismos
términos que nuestra explicación en el caso del TR y que no hace al
caso repetir ahora.
Más importante resulta recalcar la diferencia entre nuestros
resultados y los encontrados por Alguacil et al. (2013) que
encontraron un claro efecto de la congruencia previa en el efecto
Stroop tanto en la amplitud del P1 como en la del N170; en ambos
casos un ensayo previo incongruente modificaba el efecto Stroop.
Este efecto de la congruencia previa en el efecto Stroop, presentó la
forma de una reducción de la positividad del P1 y de un incremento
en la negatividad del N170 para los ensayos incongruentes respecto
a los congruentes cuando aparecían precedidos por un ensayo
incongruente, lo que añade complicación al punto, discutido
anteriormente, sobre la forma que la influencia de la congruencia
previa se manifiesta en los componentes negativos de los
CAPÍTULO 5
200
potenciales corticales evocados. Por otra parte, la principal
diferencia entre nuestro experimento y el de Alguacil et al. reside en
el tipo de estímulos y en la tarea utilizada. Alguacil et al., utilizaron
palabras en una tarea Eriksen de flancos, mientras que nosotros
hemos utilizado caras en una tarea Stroop. La razón por la que estas
diferencias producen resultados discrepantes en los componentes
tempranos no es evidente y necesita más investigación.
Finalmente merece la pena resaltar los resultados
encontrados en el componente P2 en relación con el porcentaje de
incongruencia. Tanto en el primero como en el tercer experimento,
en los que las palabras previas al objetivo fueron enmascaradas y
las máscaras se asociaron a diferentes porcentajes de incongruencia,
obtuvimos un efecto principal de esta variable de forma que la
amplitud del P2 para la condición de 66% de ensayos incongruentes
presentó mayor amplitud que para la condición de 33% de ensayos
incongruentes. Este resultado parece robusto e indica que, a lo
largo de la tarea, la asociación entre las diferentes máscaras y sus
correspondientes porcentajes de incongruencia se va fortaleciendo a
pesar de que la relación de congruencia o incongruencia entre el
estímulo previo y el objetivo no es procesada conscientemente. La
sensibilidad del componente P2 a esta asociación sugiere que este
componente puede constituir un índice adecuado de aprendizaje
implícito.
El procesamiento no consciente
Quizás la aportación más novedosa de esta tesis doctoral
reside en el hecho de que los efectos de adaptación al conflicto, que
DISCUSIÓN GENERAL
201
hemos comentado en el apartado anterior, se han obtenido en
condiciones de enmascaramiento de los estímulos previos.
Mediante una cuidadosa sesión previa de cálculo de umbral, se
estableció para cada participante el tiempo de presentación del
estímulo previo que lo hacía imperceptible de forma consciente. De
hecho, los tiempos que se determinaron mediante este
procedimiento fueron similares a los utilizados en otros
experimentos interesados en estudiar el procesamiento no
consciente mediante enmascaramiento (Daza, M. T., Ortells, J. J., &
Fox, E. (2002); Greenwald, Klinger, & Liu, 1989; Klapp, 2007,
Ortells, Marí-Beffa, & Plaza-Ayllón, 2002). Además, al final de la
sesión experimental, se preguntó a los participantes si habían sido
conscientes de alguna palabra o de la relación entre las máscaras y
el porcentaje de incongruencia. En ambos casos las respuestas
fueron negativas.
El procedimiento que utilizamos fue un procedimiento
descendiente encaminado a establecer el umbral subjetivo de
consciencia, es decir el tiempo de presentación del estímulo previo
en el que el participante comienza a decir de forma consistente que
no percibe la palabra enmascarada. Puede decirse, por tanto, que
nuestro procedimiento estableció el umbral subjetivo de consciencia
de la palabra y que el valor de nuestros resultados encuentra en este
hecho uno de los límites de su extensión. El estudio de estos
mismos problemas en condiciones de umbral objetivo de
consciencia de las palabras es tarea para futuras investigaciones.
No obstante el valor de nuestra aportación no merece ser
subestimado por varias razones.
CAPÍTULO 5
202
En primer lugar, el umbral subjetivo tiene una extraordinaria
importancia para el funcionamiento cotidiano de las personas. Esta
afirmación está avalada por numerosos síndromes
neuropsicológicos, como la “vista ciega” (blindsight) (Weiskrantz,
2009), y la heminegligencia espacial (spatial neglect) (Driver &
Vuilleumier, 2001), que no sólo han mostrado la existencia de
procesamiento cognitivo en el escotoma o en la mitad del espacio
visual afectado, sino que las personas que padecen este tipo de
problemas son inconscientes de los estímulos que provocan ese
procesamiento e incapaces de iniciar una acción intencional
dirigida a los mismos. Precisamente en estas situaciones, que
podríamos calificar de umbral subjetivo de consciencia, lo que más
interesa a la investigación neuropsicológica es conocer qué y cuánto
procesamiento cognitivo puede tener lugar en las zonas del cerebro
afectadas con el fin de poder plantear estrategias terapéuticas. De
forma parecida, en la psicología experimental parece cada vez más
urgente investigar el tipo de procesamiento que puede darse en
situaciones de umbral subjetivo para poder valorar, entre otros
temas, hasta qué punto control y consciencia son procesos que van
necesariamente asociados o no. Nuestros resultados han mostrado
de forma clara que no están necesariamente asociados.
En segundo lugar, si tomamos como referencia la
clasificación de formas de procesamiento no consciente propuesta
por Dehaene, Changeux, Naccache, Sackur, & Sergent, (2006), el
procesamiento inducido en nuestros experimentos por el
enmascaramiento de las palabras previas, puede clasificarse de
subliminal, ya que no se evitó que los participantes atendieran a los
DISCUSIÓN GENERAL
203
estímulos presentados (máscara y objetivo) sino que solamente se
impidió el procesamiento de las palabras mediante
enmascaramiento. Además cuidamos que los participantes
mantuvieran la expectativa de la presencia de una palabra entre las
máscaras. En estas condiciones, Dehaene et al. (2006) han
caracterizado el procesamiento subliminal que tiene lugar de la
siguiente forma:
No puede comunicarse mediante el lenguaje. Esa
característica se cumple en nuestros experimentos ya que los
participantes expresamente dijeron que no habían visto las palabras
enmascaradas ni captado la relación entre máscaras y porcentaje de
incongruencia.
El nivel de procesamiento depende de la atención y
de la expectativa (set) de tarea. Como ya hemos dicho
anteriormente, nuestra situación experimental se planteó
precisamente para maximizar el procesamiento de las palabras
enmascaradas sin que llegaran a ser conscientemente percibidas.
La activación neuronal puede alcanzar el nivel
semántico. Como se sigue del punto anterior, buscamos
expresamente que esto fuera posible y los resultados han
demostrado que lo conseguimos.
La facilitación (priming) que se consigue es de corta
duración. En nuestros experimentos la asincronía entre la
presentación de la segunda máscara y el estímulo objetivo (SOA)
fue de un segundo, y el SOA entre comienzo de la palabra previa y
el objetivo, superior a un segundo. Este intervalo no puede
considerarse corto, por lo que nuestros resultados muestran que la
CAPÍTULO 5
204
facilitación semántica que estímulos subliminares pueden llegar a
producir no siempre es de corta duración.
No existe una actividad frontoparietal durable.
Nuestros resultados demuestran lo contrario. Como sugieren los
mapas topográficos y se puede ver en las ondas correspondientes a
los distintos potenciales corticales analizados en los experimentos
primero y tercero, tanto el componente N2 frontal como el P3
centroparietal presentan una duración semejante a la que suele
encontrarse en condiciones de procesamiento consciente de los
estímulos. Si comparamos los resultados encontrados en los
experimentos en los que se han enmascarado las palabras previas
con el segundo experimento, en el que no existió enmascaramiento
de ningún tipo, es precisamente en el primero y tercero en los que
aparecieron efectos significativos en un N2 frontal, mientras que la
topografía del N2 en el segundo experimento no fue frontal sino
posterior. Por otra parte, la topografía y duración del componente
P3 no apareció diferente entre los tres experimentos. De hecho, si
comparamos el experimento segundo con el tercero, que presentan
grandes similitudes pero en este último se enmascararon las
palabras previas y en el segundo las palabras se pudieron procesar
conscientemente, sorprende la sensibilidad de los potenciales
corticales en el experimento de enmascaramiento. Aparecieron
efectos de latencia que no estuvieron presentes en el segundo
experimento, e interacciones de segundo y tercer orden que
tampoco estuvieron presentes en el segundo experimento. Mas
bien, esta comparación de experimentos sugiere que el hecho de ser
consciente de las palabras disminuye la sensibilidad de los
DISCUSIÓN GENERAL
205
potenciales evocados por la relación entre las palabras no
enmascaradas y los estímulos objetivo.
Finalmente, en relación con el CSPCE se ha argumentado
que para que se produzca el efecto es necesaria la representación
consciente de todos los elementos que intervienen en el conflicto: el
estímulo previo (prime), el objetivo y el contexto (Kunde, Reuss, &
Kiessel (2012). Nuestros resultados, tanto en el primer experimento
como en el tercero han demostrado que estas restricciones no son
necesarias y que el CSPCE puede tener lugar sin representación
consciente del estímulo previo (prime) y sin consciencia de la
relación entre el contexto y el porcentaje de incongruencia asociado
al mismo.
Conclusiones
Durante mucho tiempo, la distinción entre procesos automáticos
y procesos controlados (Posner & Snyder, 1975; Schneider &
Shiffrin, 1977; Shiffring & Schneider, 1977) ha dirigido la
investigación psicológica y ha sido ampliamente aceptada como
marco general de referencia de dicha investigación. Esta división
ha funcionado como una clasificación disjunta en la que cada una
de las dos clases de procesos presentaba unas características
definitorias. Por ejemplo, los procesos automáticos se consideraban
rápidos, dirigidos por los estímulos, libres de la necesidad de
recursos atencionales e inconscientes. A su vez, los procesos de
control se consideraban lentos, dependientes de la voluntad del
sujeto, necesitados de recursos atencionales y conscientes. En un
CAPÍTULO 5
206
contexto teórico como este, resulta una contradicción en los propios
términos hablar de control automático.
Nuestros resultados ponen seriamente en tela de juicio esta
distinción de los procesos psicológicos. En este punto nuestra crítica
se une a la de autores que previamente han puesto en guardia sobre
lo inadecuada que puede resultar esta división (Bugg & Hutchison,
2013; Crump, Vaquero, & Milliken, 2008; Shedden et al., 2012). A
lo largo de esta investigación hemos visto que la consciencia no es
necesaria para el funcionamiento de los mecanismos de control y
que los mecanismos de control pueden deberse a procesos de
aprendizaje, que una vez establecidos, se ponen en funcionamiento
ante la presencia de un estímulo desencadenante. Concretamente,
hemos comprobado que la asociación de una determinada máscara
con un determinado porcentaje de incongruencia acaba
determinando el contexto que da lugar al CSPCE.
Nuestros resultados encuentran un marco de referencia teórico
más adecuado en modelos de control como el propuesto por
Norman y Shallice (1986; Shallice & Cooper, 2011) que distinguen
entre dos mecanismos diferentes de control. Uno de ellos, que los
autores denominan contention scheduling, está constituido por un
conjunto de esquemas generados a partir de la experiencia con
situaciones cotidianas y/o repetitivas. Estos mecanismos de control
son fruto del aprendizaje y del conjunto de asociaciones que los
esquemas de control establecen con diferentes contextos. Estos
esquemas de control mantienen entre sí relaciones de activación e
inhibición, de forma que el esquema de control dominante en una
DISCUSIÓN GENERAL
207
situación es el que logra mayor grado de activación en esa
situación. El segundo mecanismos de control es el Sistema
Atencional Supervisor que es el responsable del control consciente
y que entra en juego en situaciones en las que los esquemas de
control no pueden hacerse cargo de la situación, bien porque no
existen, como ocurre en situaciones nuevas, bien porque exista
competición entre los esquemas, como ocurre en una tarea Stroop,
bien porque haya que corregir errores, o porque las situaciones sean
difíciles o peligrosas. Estos dos mecanismos no son una
clasificación de procesos sino dos tipos diferentes de mecanismos
de control que se comunican entre sí de forma flexible. En esta
investigación hemos puesto de manifiesto la importancia de
mecanismos que controlan el conflicto, tanto cognitivo como
emocional y que son el resultado de procesos asociativos ligados al
contexto o a procesos ligados a la secuencia de los ensayos. Ambos
tipos de procesos hemos visto que pueden ocurrir en situaciones en
que las personas no son conscientes de una parte de los factores que
controlan la situación. No obstante debe tenerse muy en cuenta que
en todos los experimentos hemos mantenido constante la atención a
los estímulos presentados y hemos mantenido activa la expectativa
de las relaciones entre el estímulo previo y el objetivo (Kiefer, M.,
Adams, S. C., & Zovko, M. (2012). Como conclusión final cabe
subrayar que la asociación necesaria entre control y consciencia,
que ha sido defendida por importantes investigadores (Botvinick,
Braver, Barch, Carter, & Cohen, 2001) debe ser definitivamente
abandonada.
REFERENCIAS
Alguacil, S., Tudela, P. and Ruz, M. (2013). Cognitive and affective
control in a flanker word task: Common and dissociable brain
mechanisms. Neuropsychologia, 51(9), 1663-1672.
Anderson, J., (1982). Acquisition of cognitive skill. Psychological
Review, 89(4), 369-406
Ashton-Jones, G. & Cohen, J. D. (2005). An integrative theory of
locus coeruleus-norepinephrine function: adaptive gain and
optimal performance. Annual Review of Neuroscience, 28,
403-450.
Atalay, N. B. & Misirlisoy, M. (2012). Can contingency learning
alone account for item-specific control? Evidence from within-
and between-language ISPC effects. Journal of Experimental
Psychology: Learning, Memory, and Cognition, 38, 1578-
1590.
Barch, D. M., Braver, T. S., Sabb, F. W., & Noll, D. C. (2000).
Anterior cingulate and the monitoring of response conflict:
Evidence from an fMRI study of overt verb generation.
Journal of Cognitive Neuroscience, 12, 298–309.
Becker, E., Rinck, M., Margraf, J. & Roth, W. (2001). The
emotional Stroop effect in anxiety disorders. General
emotionality or disorder specificity?. Anxiety Disorders, 15,
147-159.
REFERENCIAS
212
Bench, C. J., Frith, C. D., Grasby, P. M., Friston, K. J., Pauleso, E,
Frackowiak, R. S. J. & Dolan, R. j. (1993). Investigations of
the functional anatomy of attention using the stroop test.
Neuropsychologia, 31(9), 907-922.
Bishop, S., Duncan, J., Brett, M. & Laerence, A. (2004). prefrontal
cortical function and anxiety: controlling attention to threat-
related stimuli. Nature Neuroscience, 7, 184-188.
Blais, C., & Bunge, S. (2010). Behavioral and neural evidence for
item-specific performance monitoring. Journal of Cognitive
Neuroscience, 22(12), 2758–67.
Blais, C., Harris, M. B., Guerrero, J. V, & Bunge, S. A. (2012).
Rethinking the role of automaticity in cognitive control.
Quarterly Journal of Experimental Psychology (2006), 65(2),
268–76.
Blais, C., Robidoux, S., Risko, E. F., & Besner, D. (2007). Item-
specific adaptation and the conflict-monitoring hypothesis: A
computational model. Psychological Review, 114(4), 1076-
1086.
Blais, C., Tudela, P.,& Bunge, S. (2007) Proportion congruency
and the neural correlates of consciousness. Paper presented at
the Annual meeting of the Society for Neuroscience, San
Diego, CA, USA.
REFERENCIAS
213
Balkin, T. J., Braun, A. R., Wesensten, N. J., Jeffries, K., Varga,
M., Baldwin, Belenky, G., & Herscovitch, P. (2002). The
process of awakening: a PET study of regional brain activity
patterns mediating the reestablishment of alertness and
consciousness. Brain, 125(10), 2308-2319.
Botvinick, M. M., Braver, T. S., Barch, D. M., Carter, C. S., &
Cohen, J. D. (2001). Conflict monitoring and cognitive control.
Psychological Review, 108(3), 624–52.
Botvinick, M. M., Cohen, J. D., & Carter, C. S. (2004). Conflict
monitoring and anterior cingulate cortex: an update. Trends in
Cognitive Sciences, 8(12), 539–546.
Braver, T.S., Gray, J.R., & Burgess, G.C. (2007). Explaining the
many varieties of working memory variation: Dual
mechanisms of cognitive control. In A.R.A. Conway, C.
Jarrold, M.J. Kane, A. Miyake, & J. Towse (Eds.), Variation in
working memory (pp. 76–106). New York: Oxford University
Press.
Bugg, J. M. & Crump, M. J. C. (2012). In support of a distinction
between voluntary and stimulus-driven control: A review of
the literature on proportion congruent effects. Frontiers in
psychology, 3, article 367.
REFERENCIAS
214
Bugg, J. M. & Hutchison, K. A. (2013). Converging evidence for
control of color-word Stroop interference at the item level.
Journal of Experimental Psychology: Human Perception and
Performance, 39(2), 433-449. doi: 10.1037/a0029145.
Bugg, J. M., Jacoby, L. L., & Chanani, S. (2011). Why it is too
early to lose control in accounts of item-specific proportion
congruency effects. Jour- nal of Experimental Psychology:
Human Perception and Performance, 37, 844–859.
doi:10.1037/a0019957
Bush, G., Luu, P., & Posner, M. I. (2000). Cognitive and emotional
influences in anterior cingulate cortex. Trends in Cognitive
Sciences, 4(6), 215–222.
Carreti, L., Martin-Loeches, M., Hinojosa, JA., & Mercado, F.
(2001). Emotion and attention interaction studied through
event-related potentials. Journal of Cognitive Neuroscience,
13, 1109-1128.
Cañadas, E., Rodríguez-Bailón, R., Milliken, B., & Lupiáñez, J.
(2012). Social categories as a context for the allocation of
attentional control. Journal of Experimental Psychology,
142(3), 934-943.
Carter, C. S., Cohen, J. D., & Mintun, M. (1995). Functional
anatomy of selective attention in normals and schizophrenia.
Biological Psychiatry,37, 9, 638-639
REFERENCIAS
215
Casey, B. J., Castellanos, F. X., Giedd, J. N., Marsh, W. L.,
Hamburger, S. D., Anne, B., et al. (1997). Implication of right
frontostriatal circuitry in response inhibition and attention-
deficit/ hyperactivity disorder. Journal of the American
Academy of Child and Adolescent Psychiatry, 36, 374–383.
Cheesman, J., & Merikle, P. M. (1986). Distinguishing conscious
from unconscious perceptual processes. Canadian Journal of
Psychology, 40, 343-367.
Clayson, P. & Larson, M. (2011). Effects of repetition priming on
electrophysiological and behavioral indices of conflict
adaptation and cognitive control. Psychophysiology, 48(12),
1621-1630.
Clayson, P. E., & Larson, M. J. (2011). Conflict adaptation and
sequential trial effects: support for the conflict monitoring
theory. Neuropsychologia, 49(7), 1953–61.
doi:10.1016/j.neuropsychologia.2011.03.023
Compton, R. (2003). The interface between emotion and attention:
A review of evidence from Psychology and Neuroscience.
Behavioral and Cognitive Neuroscience Reviews, 6(2), 115-
129.
Corballis, P. M. and Gratton, G. (2003). Independent control of
processing strategies for different locations in the visual field.
Biological Psychology, 64, 191-209
REFERENCIAS
216
Corbetta, M., Akbudak, E., Conturo, T. E., Snyder, A. Z., Ollinger,
J. M., Drury, H. A., Linenweber, M. R., Petersen, S. E.,
Raichle, M. E., Van-Essen, D. C. & Shulman, G. L. (1998) A
common network of functional areas for attention and eye
movements. Neuron 21(4):761–73.
Cox, W., Hogan, L., Kristian, M. & Race, J. (2002). Alcohol
attentional bias as predictor of alcohol abusers´treatment
outcome. Drug and Alcohol Dependence, 68, 237-243.
Crump, M. J. C., Gong, Z., & Milliken, B. (2006). The context-
specific proportion congruent Stroop effect: Location as a
contextual cue. Psychonomic Bulletin & Review, 13, 316–321.
Crump, M. J. C., & Milliken, B. (2009). The flexibility of context-
specific control: Evidence for context-driven generalization of
item-specific control settings. Quarterly Journal of
Experimental Psychology, 62, 1523-1532.
Crump, M. J. C., Vaquero, J. M. M., & Milliken, B. (2008).
Context-specific learning and control: the roles of awareness,
task relevance, and relative salience. Consciousness and
Cognition, 17(1), 22–36.
Daza, M. T., Ortells, J. J., & Fox, E. (2002). Perception without
awareness: Further evidence from a Stroop priming task.
Perception & Psychophysics, 64(8), 1316–1324.
REFERENCIAS
217
Dehaene, S., Artiges, E., Naccache, L., Martelli, C., Viard, A.,
Schurhoff, F., Recasens, C., Martinot, M. L., Leboyer, M., &
Martinot, J. L. (2003a). Conscious and subliminal conflicts in
normal subjects and patients with schizophrenia: The role of
the anterior cingulate. Proc. Natl. Acad. Sci. USA 100, 13722–
13727.
Dehaene, S. & Changeux, J. P. (2011). Experimental and
Theoretical Approaches to Conscious Processing. Neuron,
70(2), 200-227.
Dehaen, S., Changeoux, J. P., Naccache, L., Sackur, J & Sergent, C.
(2006). Conscious, preconscious, and subliminal processing: a
testable taxonomy. Trends in Cognitive Sciences, 10, 204-211
Dehaene, S., Posner, M. I. & Tucker, D.M.. (1994). Localization of
a neural system for error detection and compensation.
Psycholical. Science, 5, 303–305.
Dehaene, S., Sergent, C., and Changeux, J. P. (2003b). A neuronal
network model linking subjective reports and objective
physiological data during conscious perception. Proc. Natl.
Acad. Sci. USA 100, 8520–8525.
Desender, K., & Van den Bussche, E. (2012). Is consciousness
necessary for conflict adaptation? A state of the art. Frontiers
in Human Neuroscience, 6, 3.
REFERENCIAS
218
Desimone, R., & Duncan, J. (1995). Neural mechanisms of
selective visual attention. Annual Review of Neuroscience, 18,
193–222.
Donchin, E. & Coles, M. G. H. (1988). Is the P300 component a
manifestation of context updating? Behavioural and Brain
Sciences, 11, 357-374.
Driver, J. & Vuilleumier, P. (2001). Perceptual awareness and its
loss in unilateral neglect and extinction. Cognition, 79(1-2),
39-88.
Egner, T. (2008). Multiple conflict-driven control mechanisms in
the human brain. Trends in Cognitive Sciences, 12(10), 374–
80. doi:10.1016/j.tics.2008.07.001
Egner, T., Etkin, A., Gale, S., & Hirsch, J. (2008). Dissociable
neural systems resolve conflict from emotional versus non-
emotional distracters. Cerebral Cortex, 18(6), 1475–84.
Egner, T., & Hirsch, J. (2005). Cognitive control mechanisms
resolve conflict through cortical amplification of task-relevant
information. Nature Neuroscience, 8(12), 1784–90.
doi:10.1038/nn1594
Ekman, P., & Friesen, W. V. (1976). Measuring facial movement.
Environmental Psychology and Nonverbal Behavior, 1(1), 56–
75.
REFERENCIAS
219
Eriksen, B. and Eriksen, C. (1974). Effects of noise letters upon the
identification of a target letter in a nonsearch task. Perception
& Psychophysics, 16(1), 143-149.
Etkin, A., Egner, T., Peraza, D., Kandel, E., & Hirsch, J. (2006).
Resolving emotional conflict: a role for the rostral anterior
cingulate cortex in modulating activity in the amygdala.
Neuron, 51(6), 871–82.
Falkenstein, M., Hoormann, J., & Hohnsbein, J. (1999). ERP
components in Go/Nogo tasks and their relation to inhibition.
Acta Psychologica, 101(2-3), 267-291
Fan, J., Cu, X., Guise, K., Liu, X., Fossella, J., Wang, H. & Posner,
M. I. (2009). Testing the behavioral interaction and integration
of attentional networks. Brain and Cognition, 70(2), 209-220.
Fan, J., & Posner, M. (2004). Human attentional networks.
Psychiatrische Praxis, 31 Suppl 2, S210–4. doi:10.1055/s-
2004-828484
Fernandez-Duque, D., Posner, M.I., 1997. Relating the mechanisms
of orienting and alerting. Neuropsychologia, 35, 477–486.
Folstein, J. R., & Van Petten, C. (2008). Influence of cognitive
control and mismatch on the N2 component of the ERP: a
review. Psychophysiology, 45(1), 152–70.
REFERENCIAS
220
Friston KJ, Frith CD, Liddle PF, Frackowiak RSJ (1991):
Comparing functional (PET) images: The assessment of
significant change. Journal or Cerebral Blood Flow &
Metabolism, 11, 690-699.
Frünholz, S., Godde, B., Finke, M. & Herrmann, M. J. (2011).
Spatio-temporal brain dynamics in a combined stimulus-
stimulus and stimulus-response conflict task. Neuroimage, 54,
622-634
Funes, M. J., Lupiáñez, J., & Humphreys, G. (2010). Analyzing the
generality of conflict adaptation effects. Journal of
Experimental psychology: Human Perception and
Performance, 36(1), 147-161.
Gehring, W. J., Coles, M. G. H., Meyer, D. E., & Donchin, E.
(1990). The error-related negativity: An event-related brain
potential accompanying errors [Abstract]. Psychophysiology,
27, S34.
Gratton, G., Coles, M. & Donchin E. (1992). Journal of
Experimental psychology: General, 121(4), 480-506.
Greenwald, A. G., Klinger, M. R., & Liu, T. J. (1989). Unconscious
processing of dichoptically masked words. Memory &
Cognition, 17(1), 35–47.
REFERENCIAS
221
Hampton, A. N. & O´Doherty, J. P. (2007) Decoding the neural
substrates of reward-related decision making with functional
MRI. Proceeding of the National Academy of Science of the
United States of America, 104(4), 1377–1382.
Heil, M., Osman, A., Wiegelmann, J., Rolke, B., & Hennighausen,
E. (n.d.). N200 in the Eriksen-task: Inhibitory executive
process?
Heinemann, A., Kunde, W., & Kiesel, A. (2009). Context-specific
prime-congruency effects: on the role of conscious stimulus
representations for cognitive control. Consciousness and
Cognition, 18(4), 966–76.
Henik, A., Bibi, U. Yanai, M. & Tzelgov, J. (1997) The Stroop
effect is largest during first trials. Abstracts of the
Psychonomic Society, 2, 57.
Hillyard, S. a, Vogel, E. K., & Luck, S. J. (1998). Sensory gain
control (amplification) as a mechanism of selective attention:
electrophysiological and neuroimaging evidence.
Philosophical Transactions of the Royal Society of London.
Series B, Biological Sciences, 353(1373), 1257–70.
Hohnsbein, J, Falkenstein, M, & Hoormann, J. (1995). Effects of
Attention and Time-Pressure on P300 Subcomponents and
Implications for Mental Workload Research. Biological
Psychology, 40(1–2), 73–81. Special Issue: EEG in Basic and
REFERENCIAS
222
Hommel, B., Proctor, R. & Vu, KPL. (2004). A feature-integration
account of sequential effects in the Simon task. Psychological
Research, 68(1), 1-17.
Jacoby, L.L., Lindsay, D. S., & Hessels, S. (2003). Item-specific
control of automatic processes: Stroop process dissociations.
Psychonomic Bulletin & Review, 10, 638-644.
Kahneman, D. (1973) Attention and Effort. Englewood Cliffs, N.
J.: Prentice-Hall.
Kane, M. J., & Engle, R. W. (2003). Working-memory capacity and
the control of attention: The contributions of goal neglect,
response competition, and task set to Stroop interference.
Journal of Experimental Psychology: General, 132, 47-70.
Kanske, P., & Kotz, S. (2010). Modulation of early conflict
processing: N200 responses to emotional words in a flanker
task. Neuropsychologia, 48(12), 3661–3664.
Kanske, P. & Kotz, S. (2011). Attentional orienting towards
emotion: P2 and N400 ERP effects. Neuropsychologia, 49,
3121-3129.
Kerns, J. G., Cohen, J. D., MacDonald, A. W., Cho, R. Y., Stenger,
V. A., & Carter, C. S. (2004). Anterior cingulate conflict
monitoring and adjustments in control. Science, 303(5660),
1023–1026.
REFERENCIAS
223
Kiehl, K. A., Liddle, P. F., & Hopfinger, J. B. (2000). Error
processing and the rostral anterior cingulate: An event-related
fMRI study. Psychophysiology, 37, 216–223.
Kiefer, M., Adams, S. C. & Zovko, M. (2012). Attentional
sensitization of unconscious visual processing: Top-down
influences on masked priming. Advances in Cognitive
Psychology, 8(1), 50-61.
Klapp, S. T. (2007). Nonconscious control mimics a purposeful
strategy: Strength of Stroop-like interference is automatically
modulated by proportion of compatible trials. Journal of
Experimental Psychology: Human Perception and
Performance, 33(6), 1366-1376.
Kopp, B., Rist, F., & Mattler, U. (1996). N200 in the flanker task as
a neurobehavioral tool for investigating executive control.
Psychophysiology, 33(3), 282–294.
Kornblum, S., Hasbroucq, T., & Osman, a. (1990). Dimensional
overlap: cognitive basis for stimulus-response compatibility--a
model and taxonomy. Psychological Review, 97(2), 253–270.
Krug, M. K., & Carter, C. S. (2012). Proactive and reactive control
during emotional interference and its relationship to trait
anxiety. Brain Research, 1481, 13–36.
REFERENCIAS
224
Kunde, W., Reuss, H., & Kiessel, A. (2012). Consciousness and
cognitive control. Advances in Cognitive Psychology, 8(1), 9-
18.
Laming DRJ. 1968. Information Theory of Choice-Reaction Times.
London: Academic press.
Lehle, C., & Hübner, R. (2008). On-the-fly adaptation of selectivity
in the flanker task. Psychonomic Bulletin & Review, 15(4),
814–818.
Lindsay, D. S., & Jacoby, L. L. (1994). Stroop process
dissociations: The relationship between facilitation and
interference. Journal of Experimental Psychology. Human
Perception and Performance, 20, 219–234.
Logan, G. D. (1980). Attention and automaticity in Stroop and
priming tasks: Theory and data. Cognitive Psychology, 12,
523-553.
Logan, G. (1988) Toward an instance theory of automatization.
Psychological Review 95, 492–527
Logan, G. D., & Zbrodoff, N. J. (1979). When it helps to be misled:
Facilitative effects of increasing the frequency of conflicting
stimuli in a Stroop-like task. Memory & Cognition, 7, 166–
174.
REFERENCIAS
225
Logan, G. D., Zbrodoff, N. J., & Williamson, J. (1984). Strategies
in the color-word Stroop task. Bulletin of the Psychonomic
Society, 22(2), 135–138.
Lowe, D.G. & Mitterer, J. O. (1982). Selective and divided
attention in a Stroop task. Canadian Journal of
Psychology/Revue canadienne de psychologie, 36(4), 684-700.
Luck, SJ. An introduction to the Event-Related Potential Technique.
MIT Press; Cambridge, MA (2005).
MacLeod, C.M. (1991). Half a century of research on the Stroop
effect: an integrative review. Psychological Bulletin, 109, 163–
203.
Marrocco, R.T., Davidson, M.C., 1998. Neurochemistry of
attention. In: Parasuraman, R. (Ed.), The Attentive Brain. MIT,
Cambridge, MA, pp. 35 – 50.
Martens, U., Ansorge, U., & Kiefer, M. (2011). Controlling the
unconscious: Attentional task sets modulate subliminal
semantic and visuo-motor processes differentially.
Psychological Science, 22, 282–291
McDonald, A. W. (2000). Dissociating the Role of the Dorsolateral
Prefrontal and Anterior Cingulate Cortex in Cognitive Control.
Science, 288(5472), 1835–1838.
REFERENCIAS
226
McDonald, A.W., Cohen, J.D., Strenger, V. A., & Carter, C. S.
(2000). Dissociating the role of the dorsolateral prefrontal
cortex and anterior cingulate cortex in cognitive control.
Science, 288(9), 1835-1838.
Menon, V., Adleman, N. E., White, C. D., Glover, G. H., & Reiss,
A. L. (2001). Error-related brain activation during a go/no-go
response inhibition task. Human Brain Mapping, 12, 131–143
Merikle, P. M. & Joordens, S. (1997). Parallels between perception
without attention and perception without awareness.
Consciousness and Cognition 6(2-3), 219-236.
Miller, E. K., & Cohen, J. D. (2001). An integrative theory of
prefrontal cortex function. Annual Review of Neuroscience, 24,
167–202.
Montalan, B., Caharel, S., Personnaz, B., Le Dantec, C., Germain,
R., Bernard, C., Rebaï, M. (2008). Sensitivity of N170 and late
positive components to social categorization and emotional
valence. Brain Research, 1233, 120–128.
Nieuwinhuis, S., Stins, J., Posthuma, D., Polderman, T., Boonsma,
D. and de Geus, J. (2006). Accounting for sequential trial
effects in the flanker task: Conflict adaptation or associative
priming? Memory & Cognition, 34(6), 1260-1272.
REFERENCIAS
227
Norman, D. A. & Shallice, T. (1986) Attention to Action. Willed
and Automatic Control of Behavior. Conciousness and Self-
Regulation, 4, 1-18.
Ortells, J. J., Marí-Beffa, P & Plaza-Ayllón, V. (2002).
Unconscious congruency priming from unpracticed words is
modulated by prime-target semantic relatedness. Journal of
Experimental Psychology: Learning, Memory, and Cognition,
39(2), 394-413.
Oschner, K., Hughes, B., Robertson, E., Cooper, J and Gabrielli, J.
(2009). Neural Systems Suporting the Control or Affective and
Cognitive Conflicts. Journal of Cognitive Neuroscience, 21(9),
1841-1854.
Panadero, A., Castellanos, M. C., and Tudela, P. (2015).
Unconscious context-specific proportion congruency effect in
a stroop-like task. Consciousness and cognition, 31, 35-45.
Pardo, J. V., Pardo, P. J., Janer, K. W. & Raichle, M. E. (1990).
The anterior cingulate cortex mediates processing selection in
the Stroop attentional conflict paradigm. Neurobiology, 87,
257-259
Perrin, F., Pernier, J., Bertrand, O., & Echallier, J. F. (1989).
Spherical splines for scalp potential and current density
mapping. Electroencephalography and Clinical
Neurophysiology, 72(2), 184–187.
REFERENCIAS
228
Petersen, S. E., Fox, P. T., Posner, M. I., Mintum, M. & Raichle, M.
E. (1988). Positron emission tomographic studies of the
cortical anatomy of single-words processing. Nature, 331, 585-
589.
Petersen, S. M. & Posner, M. I. (2012) The Attention System of the
Human Brain: 20 Years After. Annual review of neuroscience,
35, 73-89.
Polich, J. (2007). Updating P300: an integrative theory of P3a and
P3b. Clinical Neurophysiology : Official Journal of the
International Federation of Clinical Neurophysiology,
118(10), 2128–48. doi:10.1016/j.clinph.2007.04.019
Posner, M.I. (1978). Chonometric explanations of mind. Hillsdale,
NJ: Erlbaum.
Posner, M.I. (1980). Orienting of attention. Quarterly Journal of
Experimental Psychology, 32, 3-25.
Posner, M. I., & Boies, S. J. Components of attention.
Psychological Review, 1971 ,78, 391.
Posner, M. I., & Cohen, Y. (1984) Components of visual orienting.
In Attention and Performance X, H. Bouma and D. Bowhuis,
eds., pp. 531-556,
Posner, M.I. & Dehaene, S. (1994). Attentional networks. Trends
in Neuroscience, 17, 75-79.
REFERENCIAS
229
Posner, M. I., Inhoff, A. W., Friedrich, F. J. & Cohen, A. (1987)
Isolating attentional system: A cognitive-anatomical analysis.
Psychology, 15(2), 107-121.
Posner, M. I., Klein, R., Summers, J. &y Buggie, S. (1973). On the
selction of signals. Memory & Coognition, 1, 2-12.
Posner, M.I. & Petersen, S. E. (1990). The attention system of the
human brain. Annual Reviews in Neuroscience, 13, 25-42.
Posner, M. I. & Rothbart, M. K. (1991). Components of Visual
Orienting in Early Infancy: Contingency Learning,
Anticipatory Looking, and Disengaging. Journal of Cognitive
Neuroscience, 3(4), 335-344.
Posner, M. I. & Snyder, C. R. R. (1975). Attention and cognitive
control. In Information Processing and Cognition (pp. 55-85).
Hillsdale, NJ: Erlbaum
Posner, M.I & Zinder, C.R.R Facilitation and inhibition in the
processing of signals. In P. M. A. (Rabbit (Ed.) Attention and
Performace V. London: Academia Press. 1975.
Rabbitt, P. M. A. (1966). Errors and error-correction in choice-
response tasks. Journal of Experimental Psychology, 71, 264-
272.
Ridderinkhof, K. R., Ullsperger, M., Crone, E. A., & Nieuwenhuis,
S. (2004). The role of the medial frontal cortex in cognitive
control. Science (New York, N.Y.), 306(5695), 443–7.
REFERENCIAS
230
Rugg MD, & Coles MGH. (1995). ERP studies of memory. In:
Electrophysiology of Mind: Event-related Brain Potentials and
Cognition. New York: Oxford University Press, pp 132–170.
Ruz, M. & Lupiáñez, J. (2002). A review of Attentional Capture:
On it´s automaticity to endogenous control. Psicológica, 23,
283-309.
Ruz, M., Madrid, E., Lupiáñez, J., & Tudela, P. (2003). High
density ERP indices of conscious and unconscious semantic
priming. Cognitive Brain Research, 17, 3, 719-731.
Schmidt, J.R. and Besner, D. (2008) The Stroop effect: why
proportion congruent has nothing to do with congruency and
everything to do with contingency. Journal of Experimenta.
Psycholpgy: Learning, Memory and Cognition, 34, 514–523
Schmidt, J. R., Crump, M. J. C., Cheesman, J., & Besner, D. (2007).
Contingency learning without awareness: Evidence for implicit
control. Consciousness and Cognition, 16, 421- 435.
Schneider, W., & Shiffrin, R. M. (1977). Controlled and automatic
human information processing: I. Detection, search, and
attention. Psychological Review, 84, 1-66.
Scott, G. G., O’Donnell, P. J., Leuthold, H., & Sereno, S. C. (2009).
Early emotion word processing: evidence from event-related
potentials. Biological Psychology, 80(1), 95–104.
REFERENCIAS
231
Shallice, T, and Cooper, P. (2011) The organization of mind.
Oxford: Oxford University Press.
Shiffrin R.M & Schneider, W. (1977) Controlled and Automatic
Human Information Processing II. Perceptual Learning,
Automatic Attending, and a General Theory. Psychological
Review, 84(2), 127-191.
Shin, L., Whalen, P., Pitman, R., Bush, G., Macklin, M., lasko, N.
et al. (2001). An fMRI study os anterior cingulate cortex
function in postraumatic stress disorder. Biological Psychiatry,
50, 932-942.
Stroop, J.R.(1935). Studies of interference in serial verbal reactions.
Journal of Experimental psychology,18, 643-662.
Sturn, W. & Willmes, K. (2001). On the functional neuroanatomy
of intrinsic and phasic alertness. Neuroimage, 14, 78-84.
Taylor, S.F., Kornblum, s., Minoshima, S., Oliver, L. M. & Koeppe,
L. A. (1994) Changes in medial cortical blood flow with a
stimulus-response compatibility task. Neuropsychologia 32,
249–255.
Thomas, S., Jhonstone, S & Gonsalvez, C. (2007). Event-related
potentials during an Emotional Stroop task. International
Journal of Psychophysiology, 63(3), 221-231.
REFERENCIAS
232
Thompson, K. G, Biscoe, K. L, Sato, T. R. (2005). Neuronal basis
of covert spatial attention in the frontal eye field. Journal of
Neuroscience,25,9479–9487
Van Veen, V., & Carter, C. (2002). The anterior cingulate as a
conflict monitor: fMRI and ERP studies. Physiology &
Behavior, 77(4-5), 477–482.
Van Veen, V. & Carter, C. S. (2006). Conflict and Cognitive
Control in the Brain. Current Directions in Psychological
Science, 15, 237-240.
Wager, T. D., Davidson, M. L., Hughes, B. L., Lindquist, M. A., &
Ochsner, K. N. (2008). Prefrontal-subcortical pathways
mediating successful emotion regulation. Neuron, 59(6), 1037–
50. doi:10.1016/j.neuron.2008.09.006
Warburton, E., Wise, R., Price, C. J., Weiller, C., Hadar, U.,
Ramsay, S. & Frackowiak, R. S. J. (1996). Noun and verbal
retrival by normal subject: Studies with PET. Brain, 119, 159-
179.
Watts, F., McKenna, F. Sharrock, R. and Trezise, L. (1986) Colour
naming of phobia-related words. Brittish Journal of
Psychology, 77(1), 97-108.
Weiskrant, L. (2009). Blindsight: a case study spanning 35 years
and new developments. 2nd edn. Oxford University Press,
Oxford.
REFERENCIAS
233
Wendt, M. & Luna-Rodríguez, A. (2009). Conflict-Frequency
affects Flanker Interference: Role of Stimulus-Ensemble-
Specific Practice and Flanker-Response Contingencies.
Experimental Psychology, 56(3), 206-217.
West, R. & Baylis, G. C. (1998). Effects of increased response
dominance and contextual disintegration on the Stroop
interference effect in older adults. Psychology and aging,
13(2), 206-217.
Whalen, P., Bush, G., McNally, R., Wilhelm, S., McInerney, S.,
Jenike, M., & Rauch, S. (1998). The emotional counting
Stroop paradigm: a functional magnetic resonance imaging
probe of the anterior cingulate affective division. Biological
Psychiatry, 44(12), 1219–1228
Wise, R., Chollet, F., Hadar, U., Friston, K., Hoffner, E. &
Frackowiak, R. S. J. (1991). Distribution of cortical neural
networks involved in woed comprehension and word retrival.
Brain, 114(4), 1803-1817.
Wong, P. S., Bernat, E., Snodgrass, M. & Shevrin, H. (2004).
Event-related brain correlates of associative learning without
awareness. International Journal of Psychophysiology, 53(3),
217-231.
REFERENCIAS
234
Yetkin, F. Z, Hammeke, T. A., Swanson, S. J., . Morris, G. L.,
Mueller, W. M., . McAuliffe, T. L. & Haughton, V. M. (1995).
A Comparison of Functional MR Activation Patterns during
Silent and Audible Language Tasks. Amwrican Journal of
Neuroradiology, 16, 1087-1092 Zhu, H. R.,
Zhang, H. J., Wu, T. T., Luo, W. B., & Luo, Y. J. (2010). Emotional
conflict occurs at an early stage: Evidence from the emotional
face-word Stroop task. Neuroscience Letters, 478, 1–4